Retbindin Dynamically Redistributes to Preserve Retinal Flavin‐Dependent Homeostasis in Response to Riboflavin Availability and Circadian Demand Authors: Xue Zhao, Mustafa S. Makia, Gennadiy P. Moiseyev, Muna I. Naash, Muayyad R. Al‐Ubaidi Published: 2026 ABSTRACT ABSTRACTFlavin cofactors (FMN and FAD) are essential mediators of mitochondrial respiration, antioxidant defense, and redox signaling in metabolically active tissues. The neural retina exhibits one of the highest energetic and redox demands in the body, yet mechanisms governing local flavin homeostasis remain poorly defined. Retbindin (RTBDN), a retina‐specific riboflavin‐binding protein, has previously been shown to regulate total retinal flavin levels and photoreceptor survival. Here, we demonstrate that RTBDN abundance and subcellular localization are dynamically modulated by developmental stage, riboflavin availability, and the light–dark cycle. RTBDN expression emerges during outer segment maturation, peaks in early adulthood, and declines with aging in parallel with retinal flavin content. Fasting reduces retinal flavin levels without altering total RTBDN protein abundance but induces redistribution of RTBDN from the outer segment‐RPE interface toward cone‐associated domains. Riboflavin repletion restores its localization. Similarly, RTBDN redistributes around cones during the dark phase and returns to outer segment tips at light onset, independent of changes in total protein levels. These findings identify RTBDN as a spatially dynamic regulator of flavin‐dependent homeostasis that adapts to nutrient flux and circadian metabolic demand. This adaptive redistribution may represent a protective mechanism to preserve cone function under fluctuating energetic conditions and suggests that disruption of flavin homeostasis may contribute to cone degeneration. Abbreviations acetylated tubulin area under the curve connecting cilium cyclic light dark‐reared flavin adenine dinucleotide flavin adenine mononucleotide high‐pressure liquid chromatography inner segments leucine methionine outer nuclear layer outer segments phosphate‐buffered saline peanut agglutinin peripherin‐2 regular chow riboflavin‐deficient chow regions of interest retinal pigment epithelium retinal pigment epithelium 65 kDa protein retbindin riboflavin transporter deficiency Introduction Flavin cofactors, including riboflavin (vitamin B2) and its active derivatives flavin mononucleotide (FMN) and flavin adenine dinucleotide (FAD), are essential mediators of cellular redox metabolism [1, 2, 3, 4, 5, 6, 7, 8, 9]. As obligate cofactors for numerous oxidoreductases, flavins support mitochondrial oxidative phosphorylation, fatty acid β‐oxidation, tricarboxylic acid cycle activity, and antioxidant defense pathways [7, 8, 10, 11, 12, 13]. Disruption of flavin availability impairs electron transport chain function, increases oxidative stress, and contributes to neurometabolic and degenerative disease [6, 7, 10, 11, 12, 13, 14, 15, 16, 17, 18, 19]. Despite their central role in cellular redox homeostasis, mechanisms that regulate tissue‐specific flavin distribution and compartmentalization remain incompletely defined. The neural retina is among the most metabolically active tissues in the body [20, 21, 22, 23, 24, 25]. Photoreceptors require continuous ATP production to sustain phototransduction, synaptic transmission, outer segment renewal, and rapid membrane turnover [22, 23, 24, 25, 26, 27, 28, 29]. These processes depend heavily on flavin‐dependent enzymes in the mitochondria and peroxisomes, rendering photoreceptors particularly vulnerable to disruptions in flavin metabolism and redox homeostasis [7, 8, 10, 28, 30, 31, 32, 33]. Consistent with this vulnerability, clinical riboflavin deficiency is associated with visual dysfunction, and experimental models of flavin depletion demonstrate retinal degeneration [19, 34]. A notable example is ariboflavinosis caused by mutations in riboflavin transporters, referred to as riboflavin transporter deficiency (RTD), a childhood‐onset disorder characterized by progressive neuropathy, respiratory compromise, muscle weakness, sensory ataxia, deafness, and vision loss [13, 35, 36, 37, 38]. Although systemic riboflavin transport mechanisms have been characterized, the local regulatory systems that maintain flavin availability within distinct retinal compartments remain poorly understood. Retbindin (RTBDN) is a retina‐specific riboflavin‐binding protein secreted by rod photoreceptors and localized primarily at the interface between photoreceptor outer segments and the retinal pigment epithelium (RPE), a critical site for nutrient and metabolite exchange [39, 40]. A smaller fraction is also present around the inner segment [39, 40]. Sequence alignment previously demonstrated that RTBDN shares significant homology with chicken riboflavin‐binding protein [39, 41], and biochemical studies have confirmed that RTBDN binds riboflavin with high specificity [39, 40]. RTBDN plays a central role in maintaining retinal riboflavin homeostasis; genetic ablation of RTBDN results in reduced retinal flavin levels, mitochondrial dysfunction, increased oxidative stress, and progressive rod and cone degeneration, and exacerbates degeneration in models of inherited retinal disease [40, 42, 43, 44]. These findings identified RTBDN as a critical determinant of retinal flavin homeostasis and redox stability. Despite these findings, it remains unclear whether RTBDN functions primarily as a static extracellular flavin reservoir or as a dynamically regulated mediator of flavin distribution in response to metabolic demand. Retinal metabolic requirements fluctuate across development, aging, nutrient availability, and the light–dark cycle [2, 8, 9, 10, 31, 45], and the circadian clock [46, 47]. Eye‐opening and photoreceptor maturation coincide with substantial increases in oxidative metabolism [48, 49]. Aging is associated with mitochondrial decline and increased oxidative vulnerability [50]. Fasting and dietary riboflavin deficiency reduce systemic flavin availability and may challenge retinal redox balance [7, 8, 10, 34]. In addition, transitions between light and darkness impose rapid shifts in photoreceptor energy consumption and mitochondrial activity [45]. These dynamic conditions suggest that flavin‐binding capacity within the retina may require spatial and temporal adaptation to maintain metabolic resilience. Based on these observations, we hypothesized that RTBDN functions as a spatially adaptive flavin‐buffering system whose abundance and subcellular localization are regulated by developmental stage, riboflavin availability, and circadian metabolic demand. In the present study, we demonstrate that RTBDN expression parallels photoreceptor maturation and aging‐associated changes in flavin levels, whereas riboflavin deprivation and diurnal transitions induce pronounced redistribution of RTBDN without altering total protein abundance. These findings support a model in which dynamic compartmentalization of riboflavin‐binding capacity contributes to preservation of flavin‐dependent redox homeostasis in a highly oxidative neural tissue. Materials and Methods Animals All animal experiments were approved by the University of Houston Institutional Animal Care and Use Committee (IACUC) and followed NIH guidelines for the care and use of laboratory animals, as well as the guidelines from the Association for Research in Vision and Ophthalmology (ARVO). All mice were wild type (WT) and maintained on a C57BL/6N‐129S1/SvImJ background. The rd8 mutation, a naturally occurring single‐base deletion in the Crb1 gene that is known to cause retinal degeneration and can confound retinal phenotypes if present [51], was bred out. In addition, all experimental mice carried the RPE65Leu450 variant, rather than the Met450 variant. RPE65 polymorphism is known to influence visual cycle kinetics and susceptibility to light‐induced retinal damage, with the Leu450 variant associated with relatively faster chromophore regeneration and increased light sensitivity compared to the Met450 variant [52]. Both male and female mice were included in the study and exhibited comparable responses under all experimental conditions. The ages of the mice are specified where appropriate. The Rtbdn −/− mice used as primary controls were described previously [40]. Littermates were used as experimental controls. To minimize experimental bias, treatment identities were masked until completion of all analyses, including immunofluorescence and immunoblotting. The number of biological replicates is indicated for each experiment. Mice were euthanized by gradual exposure to carbon dioxide (CO2), in accordance with the AVMA Guidelines on Euthanasia. Mice were housed under standard dark–light cycles of 12‐h light/12‐h dark and light intensity of 30 lux. The light cycle starts at 7 a.m. and ends at 7 p.m. Tissue collection for the dark‐phase samples was performed at 6:00 a.m. under dark conditions, whereas light‐phase samples were collected during the light period (8:00 a.m., 11:00 a.m., 1:00 p.m., and 3:00 p.m.). Littermates were divided into two groups and assigned either fasted or fed. In the fed condition (controls), mice had ad libitum access to regular chow (RC) (AIN‐93G; Envigo, Indianapolis, IN), whereas fasted mice were deprived of food for 16 h with free access to water. Fasting was initiated at 6:00 p.m., and tissues were collected the following morning at 10:00 a.m. The timeline is described in the figure. Again, littermates were divided into two groups; one group was fed ad libitum regular chow (RC) (AIN‐93G; Envigo, Indianapolis, IN), whereas the other group was fed ad libitum a riboflavin‐deficient chow (RDC) (Riboflavin‐Deficient‐AIN‐93G; Envigo, Indianapolis, IN). Antibodies Antibodies used are summarized in Table S1. High‐Performance Liquid Chromatography (HPLC) Retinoid Profile Assay HPLC‐based retinoid profile was performed as previously described [53]. Briefly, animals were dark‐adapted overnight and euthanized under dim red light. Eyes were immediately enucleated and flash‐frozen in liquid nitrogen. Whole eyes were homogenized using a glass grinder in a lysis buffer containing 10 mM NH2OH, 50% ethanol, and 50% 2‐(N‐morpholino) ethanesulfonic acid (MES), pH 6.5. Retinoids were extracted with hexane, and the organic phase was evaporated under argon gas. Dried retinoids were resuspended in 200 μL of HPLC mobile phase (11.2% ethyl acetate, 2.0% dioxane, 1.4% octanol, 85.4% hexane). Samples were injected into an HPLC system (515 HPLC pump; Waters Corp., Milford, MA) equipped with a normal phase Lichrosphere SI‐60 column (5 μm; Alltech, Deerfield, IL) and eluted under isocratic conditions at a flow rate of 1 mL/min. Retinal Flavin Assay Micro‐extraction of flavins from biological tissue was performed as previously described [31]. Briefly, frozen tissues were homogenized using a handheld motorized pestle in 100 μL phosphate‐buffered saline (1× PBS, pH 5.1). A 30 μL aliquot of the homogenate was reserved for protein concentration measurements. The remaining homogenate was heated at 80°C for 30 s, cooled to room temperature, and treated with 10% trichloroacetic acid for 15 min to precipitate proteins. Samples were centrifuged at 10 000 × g for 10 min at 4°C, and the supernatant was carefully collected and filtered through a 0.45 μm membrane filter prior to HPLC analysis. HPLC was performed immediately after extraction using a Waters system (Water Corporation, Milford, MA, USA) to prevent alterations in flavin concentration due to repeated freeze–thaw cycles. Chromatographic separation and flavin quantification were carried out as previously described [31], using a Waters X‐Bridge C18 column (3.5 μm, 4.6 × 250 mm). Immunoblotting and Protein Quantification Retinas were collected from euthanized mice at the indicated ages, flash frozen in liquid nitrogen, and stored at −80°C until use. Frozen retinas were lysed in cold phosphate‐buffered saline (PBS, pH 7.4) containing 1% Triton X‐100 and protease inhibitors (Roche, Basel, Switzerland). Samples were homogenized using a hand‐held homogenizer (VWR, Randor, PA), sonicated with 10 pulses, and incubated at 4°C for 1 h. Homogenates were centrifuged at 4000 × g for 10 min, and supernatants were collected. Protein concentrations were determined using the Bradford Assay (Bio‐Rad, Hercules, CA). Thirty micrograms of protein were mixed with Laemmli buffer [54] containing β‐mercaptoethanol (BME), incubated at room temperature for 10 min, and then heated at 95°C for 5 min. Samples were separated on 12% SDS‐PAGE gels and subsequently transferred onto a PVDF membrane (IEVH00005; Merck Millipore) using Trans‐Blot Turbo Transfer System (Bio‐Rad, Hercules, CA). Membranes were blocked with 5% nonfat milk and incubated with specific primary antibodies (Table S1) to detect the proteins of interest. Signals were visualized using SuperSignal West Pico PLUS Chemiluminescent Substrate (Thermo Fisher Scientific, Waltham, MA, USA) and imaged with the ChemiDoc Imaging system (Version 3.0.1.14; Bio‐Rad, Hercules, CA). Quantitative analysis was performed on unsaturated bands using Image Lab software (Version 6.1; Bio‐Rad). Band intensities were normalized to β‐actin levels from the corresponding lanes of each immunoblot. Immunofluorescence Following euthanasia, mouse eyes were enucleated and incubated in modified Davidson Fixative [34] modified to 33% ethanol, 11% acetic acid, and 9% paraformaldehyde for 3–4 h. Eyes were then washed with PBS and stored in 70% ethanol. Samples were dehydrated using a STP 120 Spin Tissue Processor (Thermo Scientific, Waltham, MA, USA). After paraffinization, eyes were embedded in plastic molds (Tissue Tek Sakura; Finetek, Torrance, CA, USA) and sectioned at 10 μm thickness. Deparaffinized sections were subjected to antigen retrieval in Tris‐EDTA buffer (pH 9.0) containing 0.1% Tween‐20 for 15 min. After cooling, sections were rinsed with water and treated with 1% NaBH4 (452882‐500G; Sigma‐Aldrich) for 3 min, followed by a water rinse. Sections were then incubated in blocking buffer (PBS, pH 7.4, 2% donkey serum, 0.5% Triton X‐100, 50 mg/mL bovine serum albumin) for 1 h at room temperature. After blocking, sections were incubated overnight at 4°C with primary antibodies (Table S1), washed three times in PBS (10 min each), and incubated for 2 h with appropriate fluorescent secondary antibodies (Table S1). Sections were washed three times in PBS and counterstained with DAPI (Life Technologies, Carlsbad, CA) for 30 min. Prolong Gold antifade Mountant (Thermo Fisher Scientific, Waltham, MA, USA) was applied prior to imaging. Confocal images were acquired using a 63X oil‐immersion objective (NA1.4) on a Zeiss LSM 800 confocal/Airyscan microscope (Carl Zeiss, Oberkochen, Germany), capturing a 10‐slice stack spanning 1.71 μm, which was collapsed into a single projection image. Image processing was performed with Zen Image Analysis software (Carl Zeiss) and Photoshop (Adobe, San Jose, CA). Three‐dimensional reconstructions were generated using the 3D Create tool in ZEN 3.2 and exported in TIFF format. Signal Intensity Measurements For signal intensity measurements, retina sections were labeled for EZRIN (green) to identify the RPE microvilli and PNA was used to label cone photoreceptors. RTBDN signal intensity was measured in images acquiredwith a Zeiss LSM 800 (Zen 3.12, version 3.12.107.03000; Carl Zeiss Microscopy GmbH) using a 63× objective. In Figure 4D, RTBDN signal intensity was quantified across the entire image area (50.71 × 50.71 μm), spanning from the RPE to the photoreceptor layer, to compare fed (gray) and fasted (green) conditions. For Figure 6, images (50.71 × 50.71 μm) obtained at different time points were divided into three adjacent regions of interest (ROIs; ROI‐1–3, indicated by dashed boxes). RTBDN signal intensity was measured in each ROI from the RPE through the photoreceptor layer. In Figure 7E, a rectangular ROI (4.31 × 25.85 μm) was drawn to include a single cone labeled with PNA at the center. RTBDN signal intensity was quantified along the axis extending from the OS tips to the IS at 6:00 a.m. and 3:00 p.m. Signal intensity profiles were generated with Zeiss Profile tool and plotted with GraphPad Prism software (GraphPad Software Inc., Boston, MA, USA). Data are presented as mean ± SEM. The connecting lines/curves indicated the mean values of RTBDN signal intensity profiles, with shaded areas indicating the corresponding error bands. Area under the curve (AUC) values are indicated in each figure. Statistical Analysis Sample sizes and the numbers of experimental replicates are indicated in the figure legends. Differences between two groups were determined using an unpaired t‐test non‐parametric Mann–Whitney tests. Comparisons among multiple groups were performed using Kruskall–Wallis test with Dunn's multiple comparisons test. Statistical significance was defined as p < 0.05 (*), p < 0.01 (**), p < 0.001 (***), and p < 0.0001 (****). All analyses were conducted using GraphPad Prism software (GraphPad Software Inc., Boston, MA, USA). Results RTBDN Selectively Maintains Retinal Flavin Homeostasis Without Altering Retinoid Levels Riboflavin is the precursor for the redox cofactors FMN and FAD, which are essential for mitochondrial electron transport and multiple oxidoreductase reactions. RTBDN, a riboflavin‐binding protein expressed by rods, has been identified as a key regulator of flavin homeostasis in the retina [39, 40]. In contrast, Fan et al. [55] proposed that RTBDN may mediate light‐induced retinal damage, potentially through a role in 11‐cis‐retinal transport, suggesting that loss of RTBDN could influence retinoid metabolism and contribute to premature retinal aging. To directly evaluate whether RTBDN binds or transports 11‐cis retinal, we quantified key retinoids associated with the retinoid recycling pathway that occurs between the retinal pigment epithelium (RPE) and photoreceptor outer segments (OS) [53, 56]. Therefore, we quantified retinoid levels from the whole eye collected in Rtbdn −/− mice at P30, an age preceding detectable structural or functional degeneration [40]. Consistent with our previous findings, total retinal flavin levels were significantly reduced in Rtbdn −/− retinas (46.6%, p = 0.0286) compared to wild‐type (Rtbdn +/+, WT) controls (Figure 1A). In contrast, retinoid levels were indistinguishable between genotypes (Figure 1B), indicating that RTBDN deficiency selectively disrupts retinal flavin homeostasis without altering retinoid metabolism. These findings confirm a specific role for RTBDN in maintaining retinal riboflavin availability and flavin‐dependent redox pathways rather than participating in the visual cycle. RTBDN specifically binds riboflavin. (A) Levels of riboflavin, FMN and FAD in the retina of WT and Rtbdn −/− mice at P30. (B) Quantification of 11‐cis retinal, all‐trans retinal, and retinyl esters in whole eyes from WT and Rtbdn −/− mice collected at P30 following overnight dark adaptation. Data are presented as mean ± SD. Statistical significance was determined using the non‐parametric Mann–Whitney tests. Sample size: N = 4 for A and N = 8 for B. p values: RF: p = 0.0286 (*), FMN: p = 0.0286 (*), FAD: p = 0.0286 (*); 11‐cis retinal: p = 0.1049 (ns), all‐trans retinal: p = 0.4418 (ns), retinyl esters: p = 0.3823 (ns). Developmental Regulation of RTBDN Expression and Localization in the Retina To determine the onset of RTBDN expression in the WT retina, immunoblot analysis was performed on retinal lysates starting at postnatal day (P) 1. RTBDN was first detected at P5 and became more prominent by P7 (Figure 2A, upper, left and center panels), coinciding with outer segment (OS) elaboration, as reflected by increasing levels of PRPH2 (Figure 2A, middle, left and middle panels). RTBDN expression increased markedly after eye opening (~P13) through early adulthood, before gradually declining with age (Figure 2A, right panel; Figure 2B). This age‐dependent decline parallels a reduction in total retinal flavin content [31], suggesting that RTBDN abundance may be linked to flavin‐dependent metabolic and redox capacity in photoreceptors. RTBDN expression during retinal development. (A) Representative immunoblots showing the developmental pattern of RTBDN and PRPH2 expression in eyes/neural retinas from WT mice at the indicated ages (Eyes: P1–P10; Neural retinas: P15–P360), and P30 Rtbdn −/− retina was included as a negative control. The P5–7 portion of the blot was exposed longer to visualize the faint bands at P5 and 7 (center panel). (B) Quantification of RTBDN (blue line) and PRPH2 (black line) protein levels normalized to actin in WT mice at the indicated ages. Data are presented as mean ± SD. Sample size: N = 3 for each age. (C) Plot showing age‐dependent changes in photoreceptor outer segment length (green; left y‐axis), total flavin levels (orange; right y‐axis), and RTBDN protein levels (blue; far‐right y‐axis). Data points represent mean values at each age without the standard deviation (N = 3). The plot was generated using BioRender (assessed on February 16, 2026). (D) Retinal paraffin sections from WT mice at the indicated ages (P1–P30) immunolabeled for RTBDN (red) and EZRIN (green), with nuclei counterstained with DAPI (blue). RTBDN signal is first detected at P7. White box marks the region shown at higher magnification in the inset. IS, inner segment; ONL, outer nuclear layer; OS, outer segment. Scale bar: 10 μm (upper panel), 2 μm (lower panel) and 1 μm in the inset. To examine the relationship between RTBDN expression and photoreceptor maturation, OS length was measured across development and plotted alongside total retinal flavin levels (Figure 2C). OS length increased rapidly after birth and stabilized around P30, following retinal maturation (Figure 2C, green line). RTBDN expression rose during OS elongation, peaked shortly after eye opening, remained relatively stable, and reached a second peak in early adulthood (~P120) before declining by P360 (Figure 2C, blue line). Total retinal flavin levels followed a similar trajectory, rising during postnatal development, remaining elevated through adulthood with a second peak around P120, and declining in older animals (Figure 2C, orange line). The parallel temporal patterns of RTBDN abundance and retinal flavins confirm RTBDN's role in supporting flavin‐dependent redox reactions during periods of high metabolic demand associated with OS growth and photoreceptor maturation. Previous studies have localized RTBDN predominantly at the interface between RPE microvilli and OS tips, with minor localization around the photoreceptor inner segment (IS) region [39, 40, 42, 43]. Consistent with these reports, developmental analysis revealed that RTBDN first appears at the RPE/OS interface (Figure 2D, P7) and gradually spreads across the OS and into the IS region as the retina matures (Figure 2D, P10, P15, and P30; Figure S1). This spatial expansion mirrors the increasing metabolic and redox demands of developing photoreceptors. Collectively, these results establish a developmental window for the onset of RTBDN expression and demonstrate that its abundance and localization are dynamically regulated throughout life. Unlike PRPH2, a structural component of the OS, RTBDN expression closely tracks total retinal flavin levels, consistent with a functional role in maintaining flavin‐dependent metabolism and redox homeostasis in the developing and mature retina. Light Exposure Regulates RTBDN Expression, Flavin Levels, and Localization in the Developing Retina Because RTBDN levels increased markedly following eye opening (P15), we examined whether RTBDN expression is regulated by light exposure. Mice were reared either in complete darkness (dark‐reared; DR) or under cyclic light (CL) conditions beginning at P5, and retinas were collected at P15, P21, and P30 for analysis (Figure 3A). As shown in Figure 3B,C, immunoblot analysis demonstrated a progressive developmental upregulation trend of RTBDN protein from P15 to P30 in both groups (Figure 3B,C; Figure S2A). Although statistical significance was not reached, RTBDN levels in DR retinas at P15 displayed a trend toward lower expression (61.9%) compared with CL retinas (Figure 3B,C; Figure S2A), whereas PRPH2 protein levels were comparable under DR and CL conditions (Figure 3B,C; Figure S2B). Total retinal flavin levels mirrored this pattern (Figure 3D; Figure S2C–E), with approximately a 52% decrease in DR retinas relative to CL controls. Notably, the differences in RTBDN and total flavin levels observed at P15 were no longer detected at P21 or P30 (Figure 3C,D), suggesting that light‐dependent regulation of RTBDN occurs primarily during the developmental period surrounding eye opening. Although total flavin levels were comparable at later ages, flavin composition differed: FMN levels were significantly elevated in DR retinas relative to CL retinas, whereas FAD levels showed the opposite trend (Figure S2C). Dark rearing alters flavin levels and RTBDN levels and localization. (A) Experimental design. Mice were raised under dark (DR) or cyclic light (CL) conditions beginning at P5. Retinal tissues were collected at P15, P21, and P30 for analyses of RTBDN protein levels, total flavin content, and RTBDN localization. This schematic was generated using BioRender.com (accessed on February 16, 2026). (B) Representative immunoblots of retinal extracts probed for RTBDN and PRPH2 from DR and CL groups at P15, P21, and P30. P30 Rtbdn −/− retina was included as a negative control. (C) Quantification of RTBDN and PRPH2 protein levels normalized to actin in DR and CL groups across developmental time points. Data are presented as mean ± SD. Statistical significance was determined using the non‐parametric Mann–Whitney tests. Sample size: N = 3 for all groups. (D) Total retinal flavin levels in DR and CL groups across developmental time points. Data are presented as mean ± SD. Statistical significance was determined using an unpaired t‐test and non‐parametric Mann–Whitney tests. Sample size: N = 3 for all samples. (E–G) Representative confocal images and 3D renderings showing RTBDN localization in retinal cross‐sections from CL and DR mice at P15 and P30. (E) Brightfield (BF) and immunofluorescence labeling of RTBDN (red), EZRIN (green) and DAPI (blue). White arrowheads denote RTBDN signal within the photoreceptor layer (F). Co‐staining for RTBDN (red), acetylated tubulin (AcTUB, green), and DAPI (blue). White arrowheads indicate RTBDN signal in the periciliary region. (G) Co‐staining for RTBDN (red), AcTUB (green), PNA (white) and DAPI (blue). Yellow arrowheads designate RTBDN signal surrounding the cones. (H) Quantification of RTBDN signal intensity from confocal images comparing dark‐reared (pink) and cyclic light (blue) conditions at P15 and P30. Area under the curve (AUC) values are indicated. Data are presented as mean ± SD. Sample size: N = 6 per condition. IS, inner segments; ONL, outer nuclear layer; OS, outer segments; RPE, retinal pigment epithelium. Scale bar: 5 μm. Having established that RTBDN expression is influenced by light onset, we next assessed whether its subcellular localization is similarly affected. Retinal sections were co‐labeled for RTBDN and EZRIN, a marker of RPE microvilli. Although quantitative differences in RTBDN abundance were most pronounced at P15 (Figure 3B,C), subtle but consistent differences in localization were observed between CL and DR retinas (Figure 3E, white arrowheads). At P15, RTBDN in both CL and DR retinas was predominantly distributed around photoreceptor OSs. By P30, however, distinct localization patterns emerged despite similar overall RTBDN levels. In DR retinas, RTBDN was concentrated at the interface between the outer and inner segments (OS–IS junction), whereas in CL retinas it was distributed around OSs with enrichment at OS tips (Figure 3E, white arrowheads). To further characterize the altered localization observed in P30 DR retinas, sections were labeled with acetylated tubulin (AcTUB), a marker of the connecting cilium (CC) and proximal axoneme. In CL retinas at both P15 and P30, RTBDN localization appeared restricted to the apical region surrounding the CC. In contrast, in DR retinas at both ages, RTBDN extended toward the mid‐periciliary region (Figure 3F, white arrowheads in the 3D renderings). Because cone OSs are positioned at the base of rod OSs, we examined whether the altered labeling in DR retinas was associated with cones by staining with peanut agglutinin (PNA), which labels cone sheaths [57]. DR retinas showed reduced RTBDN signal surrounding rod OSs (Figure 3G, yellow arrowheads), whereas increased localization was observed around cone OSs (Figure 3G, yellow arrowheads). To further assess the spatial distribution of RTBDN under CL and DR conditions, RTBDN signal intensity profiles were generated across retinal layers at P15 and P30 (Figure 3H). At P15, overall RTBDN signal intensity was comparable between both conditions (Figure 3H, left panel). However, in P15 CL retinas, RTBDN signal peaked near the tips of the OS, whereas in DR retinas, the signal exhibited a broader distribution spanning the OS and IS regions (Figure 3H, left panel). By P30, the localization of RTBDN in DR retinas shifted further toward the IS compared with the CL conditions (Figure 3H, right panel). These findings suggest that dark rearing alters the subcellular distribution of RTBDN. Riboflavin Availability Regulates RTBDN Subcellular Localization Without Altering Protein Abundance Given that RTBDN is a riboflavin‐binding protein, we investigated whether dietary riboflavin availability influences RTBDN expression or localization. Mice were assigned to fasted/fed conditions. In the fed condition (Fed), mice had ad libitum access to food and water, no restriction before tissue collection. Although in the fasted condition (Fasted), mice were deprived of food for 16 h (6:00 p.m.—the following morning 10:00 a.m.) with free access to water. The following morning at 10:00 a.m., all samples were collected for immunohistochemical analysis (Figure 4A). Under fed conditions, RTBDN co‐localized with EZRIN at the interface between the RPE and OS, with additional localization in portions of the IS (Figure 4B, yellow arrowheads, Fed), consistent with previous reports [39]. In contrast, fasting induced a pronounced redistribution of RTBDN, with accumulation at the OS‐IS junction (Figure 4B, yellow arrowheads, Fasted). Because cone OSs reside at the base of rod OSs, sections were co‐labeled with PNA to assess whether RTBDN redistribution during fasting was associated with cones. As shown in Figure 4C (yellow arrowheads, Fasted), RTBDN became enriched around cone photoreceptors, evident from colocalization with PNA. Fasting alters RTBDN distribution and retinal flavin levels. (A) Experimental timeline for fed versus fasted conditions. P30 WT Mice were maintained under a controlled 12 h/12 h light/dark cycle. One group was fasted beginning at 6:00 p.m. (Fasted), whereas the control group was allowed to feed ad libitum (Fed). Retinal tissues were collected the following morning at 10:00 a.m. for analyses of RTBDN protein levels, total flavin contents, and RTBDN localization. The schematic was generated using BioRender.com (accessed on February 16, 2026). (B) Immunofluorescence labeling of retinal sections from P30 WT fed or fasted mice for RTBDN (red), EZRIN (green), and DAPI (blue). Yellow arrowheads highlight the RTBDN signal in the fed and fasted photoreceptor region. (C) Immunofluorescence labeling of retinal sections from P30 WT fed and fasted mice for RTBDN (red), PNA (white) and DAPI (blue). Right panels show higher magnification images of OS and IS tips to illustrate RTBDN distribution between fed and fasted conditions. OS: Outer segments, IS: Inner segments, ONL: Outer nuclear layer. Scale bar: 5 μm. (D) Quantification of RTBDN signal intensity from confocal images comparing fed (gray) and fasted (green) conditions. Mean values overlap within the 95% confidence interval error bands. Area under the curve (AUC) values are indicated. Data are presented as mean ± SD. Sample size: N = 6 per condition. (E) Representative immunoblots of retinal extracts from P30 mice under fed and fasted conditions probed for RTBDN, with P30 Rtbdn −/− retina included as a negative control. Quantification from three independent samples demonstrated no significant difference in RTBDN levels between fed and fasted groups. Data are presented as mean ± SD. Statistical significance was determined using the non‐parametric Mann–Whitney tests. Sample size: N = 3 for each. p values: p > 0.9999 (ns); p values: p = 0.0022 (**). (F) Total flavin levels in the retina of P30 either fed or fasted mice. Data are presented as mean ± SD. Statistical significance was determined using an unpaired t‐test and non‐parametric Mann–Whitney tests. Sample size: N = 6 for both groups. Quantification of RTBDN signal intensity across OS and IS layers confirmed that the total protein abundance (AUC) did not differ significantly between fed and fasted mice (Figure 4D). However, the spatial distribution shifted from predominantly OS tips to the OS‐IS junctions (Figure 4D), indicating that fasting alters RTBDN localization without changing its overall abundance. Immunoblot analysis further confirmed comparable RTBDN protein levels between fed and fasted groups (Figure 4E,F). Consistent with decreased dietary riboflavin, total retinal flavin levels were significantly reduced (36%, p = 0.0022) in fasted mice, relative to the fed group (Figure 4G). To distinguish whether RTBDN redistribution was driven specifically by riboflavin availability or by the feeding process per se, mice were fasted and then re‐fed for 5 h with either riboflavin‐deficient chow (RDC) or regular chow (RC). As controls, mice were fed ad libitum with RDC or RC (Figure 5A; Figure S3A). Immunoblot analysis showed that RTBDN protein levels remained unchanged across all feeding regimens (Figure 5B). In contrast, total retinal flavin and individual flavin species differed significantly among groups, particularly in fasted mice and in those refed RC (Figure 5C; Figure S3B). Feeding RC after fasting restored flavin levels to ~80% of baseline, relative to RC‐fed controls (Figure 5C). Introduction of riboflavin restores flavin levels and RTBDN localization. (A) Experimental design and timeline. P30 WT Mice were maintained under a controlled 12 h/12 h light/dark cycle. Mice were assigned to the following groups: Fasted; Fasted > Fed RDC for 5 h (7:00 a.m.–12:00 p.m.); Fasted > Fed RC for 5 h (7:00 a.m.–12:00 p.m.); Fed RDC; and Fed RC. Retinal tissues were collected at 12:00 p.m. for analyses of RTBDN protein levels, total flavin levels, and RTBDN localization. The schematic was created using BioRender.com (accessed on February 16, 2026). (B) Representative immunoblots of RTBDN in neural retinas from P30 WT mice under the indicated experimental conditions (Fasted; Fasted > Fed RDC 5 h; Fasted > Fed RC 5 h; Fed RDC; Fed RC). P30 Rtbdn −/− retina was included as a negative control. Quantification of RTBDN levels normalized to actin shows no statistically significant differences among groups. Data are presented as mean ± SD. Statistical significance was determined using the Kruskal–Wallis test with Dunn's multiple comparisons test. Sample size: N = 3 per group. (C) Total retinal flavin levels in P30 mice across experimental groups. Data are presented as mean ± SD. Statistical significance was determined using the Kruskal–Wallis test with Dunn's multiple comparisons test. Sample size: N = 6 for each experimental group. p values: p = 0.0051 (**); p < 0.0001 (****). (D) Immunofluorescence labeling of retinal sections from P30 mice in the indicated groups for RTBDN (red), EZRIN (green), and DAPI (blue). Arrowheads highlight RTBDN signal within photoreceptor layer. (E) Immunofluorescence labeling of acetylated tubulin (AcTUB) (green), PNA (white), and DAPI (blue) across the indicated groups (upper panel). 3D renderings illustrate RTBDN localization around cone photoreceptors. Arrowheads indicate RTBDN‐enriched regions surrounding cones. The raw data are presented in Figure S3. IS, inner segments; ONL, outer nuclear layer; OS, outer segments. Scale bar: 5 μm. Despite stable RTBDN protein levels, its subcellular localization was highly sensitive to flavin availability. The enrichment of RTBDN around cone photoreceptors observed in fasted mice persisted in mice fed RDC (Figure 5D,E; Figure S4), providing strong evidence that riboflavin availability, rather than feeding alone, drives the redistribution of RTBDN within photoreceptor layers. These results highlight a nutrient‐sensitive mechanism by which RTBDN dynamically responds to flavin levels, potentially ensuring localized support of flavin‐dependent redox reactions in metabolically demanding photoreceptor compartments. The Light Cycle Regulates RTBDN Localization Mice used in the experiments shown in Figures 4 and 5 were fasted during the dark phase and fed after light onset. To determine whether the light–dark cycle independently influences RTBDN behavior, immunohistochemical analysis was performed on retinal sections collected at 6:00 a.m. (dark phase), 8:00 a.m. (after light onset), 11:00 a.m., 1:00 p.m., and 3:00 p.m. (light phase) (Figure 6A–C). To quantitatively assess RTBDN distribution across the photoreceptor ISs and OSs, we applied an area‐under‐the‐curve (AUC) approach. Three independent regions of interest (ROIs) spanning from the RPE/OS interface to the apical portion of the outer nuclear layer were analyzed (Figure 6A). Representative RTBDN intensity profiles (Figure 6A, pink lines, right panel) were plotted as signal intensity versus distance from the top of the ROI, with retinal layers indicated. At 6:00 a.m., prior to light onset, RTBDN signal was predominantly concentrated in patches near the junction of the OS and IS. Following light onset, RTBDN progressively redistributed toward the OS layer, becoming enriched at OS tips between 8:00 a.m. and 3:00 p.m. (Figure 6B,C). Overlay of mean RTBDN intensity profiles revealed a clear time‐of‐day‐dependent shift in RTBDN localization toward the OS during the light phase (Figure 6C). The circadian behavior of RTBDN. (A) Example of region‐of‐interest (ROI) selection and RTBDN signal intensity analysis. Representative retinal cross‐section collected at 6:00 a.m. and stained for RTBDN (red) and DAPI (blue). Three adjacent ROIs (ROI‐1–3; dashed boxes) spanning from the RPE through the photoreceptor layer were used to quantify RTBDN signal intensity. The example intensity profile (right panel) shows signal intensity plotted against distance from the top of the ROI, with retinal layers indicated (IS, inner segments; ONL, outer nuclear layer; OS, outer segments; RPE, retinal pigment epithelium). Signal intensity profiles are presented as mean ± SEM. (B) Representative images at different time points (left to right: Brightfield; merged RTBDN (red) and EZRIN (green) immunofluorescence; RTBDN channel alone), with corresponding RTBDN signal intensity profiles shown in the right panel. Area under the curve (AUC) values with associated error and confidence intervals are indicated for each time point. Arrowheads denote RTBDN‐enriched regions within the photoreceptor layer. Scale bar: 5 μm. (C) Overlay of mean RTBDN intensity profiles across different time points (6:00 a.m., 8:00 a.m., 11:00 a.m., 1:00 p.m., 3:00 p.m.), plotted from the RPE through the photoreceptor layer, illustrating the distribution of RTBDN signal during the day. To determine whether this redistribution reflects changes in RTBDN abundance, immunoblot analysis was performed on retinal extracts collected at the same time points (Figure 7A). Despite the pronounced spatial redistribution, total RTBDN protein levels remained unchanged across the light–dark cycle. We next quantified retinal levels of riboflavin, FMN, and FAD at 6:00 a.m. (Figure 7B, pink bar) and normalized to the corresponding 3:00 p.m. levels (Figure 7B, blue dashed line, 100%). Retinal flavins displayed time‐dependent variation, with reduction in FAD (14%), FMN (22.7%), and RF (28.6%) levels at 3:00 p.m., relative to flavin levels at 6:00 a.m. (Figure 7B; Figure S5A), the time points that showed the greatest difference in RTBDN localization, consistent with the nocturnal feeding behavior of mice and suggesting that RTBDN redistribution may reflect dynamic regulation of flavin availability in response to feeding state. Relocation of RTBDN around cones in the dark. (A) Representative immunoblots of retinal extracts from P30 WT mice collected at 6:00 a.m., 8:00 a.m., 11:00 a.m., 1:00 p.m., and 3:00 p.m., probed for RTBDN. A P30 Rtbdn −/− retinal extract was included as a negative control. Quantification of RTBDN levels normalized to actin shows that total retinal RTBDN levels remain stable across the circadian cycle. Data are presented as mean ± SD. Statistical significance was determined using an ordinary one‐way ANOVA with Tukey's multiple comparisons test. Sample size: N = 3 per time point. (B) Quantification of retinal flavin levels in P30 WT retinas collected at 6:00 a.m. and normalized to corresponding 3:00 p.m. levels (blue dashed line, 100%). Data are presented as mean ± SD. (C) Representative confocal images of retinal cross‐sections collected at 6:00 a.m. and 3:00 p.m. stained for RTBDN (red), EZRIN (green), and DAPI (blue). Arrowheads indicate RTBDN‐enriched regions within the photoreceptor layer. (D) Representative images and 3D renderings showing RTBDN (red) enriched around cone photoreceptors labeled with PNA (white) and DAPI (blue) at 6:00 a.m. IS, inner segments; ONL, outer nuclear layer; OS, outer segments. Scale bar: 5 μm. (E) Representative of a single cone image taken as ROI at 6 a.m. and 3 p.m., labeled with PNA and RTBDN. Scale bar: 2μm. (F) Signal intensity profile of RTBDN measured within a 4.31 × 25.85 μm rectangle centered on cones (identified by PNA labeling) and extending from the OS/RPE interface to the IS. Two rectangular ROIs with cones in the center were selected per image. Data are presented as mean ± SEM. Sample size: N = 3 per time point. To determine whether the dark‐phase localization of RTBDN reflects increased association with cone photoreceptors, immunohistochemistry was performed using PNA to label the extracellular matrix surrounding cones. At 6:00 a.m., RTBDN signal prominently surrounded cones (Figure 7C,D, upper panels, white arrowheads; Figure S5B), whereas at 3:00 p.m. the signal was primarily concentrated at OS tips (Figure 7C,D, lower panels, white arrowheads; Figure S4). To quantify RTBDN signal intensity in cone regions, we selected individual cones within PNA‐positive areas as our ROIs at 6:00 a.m. and 3:00 p.m. (Figure 7E). AUC analysis confirmed that overall RTBDN abundance did not differ between these time points. However, at 6:00 a.m., RTBDN signal was more concentrated around the cones (Figure 7E,F, pink), whereas at 3:00 p.m., the signal was more prominent at the tips of the OS (Figure 7E,F, blue). Together, these findings demonstrate that RTBDN undergoes pronounced light–dark cycle‐dependent redistribution without changes in total protein abundance. This dynamic localization likely reflects the shifting metabolic and redox requirements of rod and cone photoreceptors across the day, consistent with a role for RTBDN in regulating flavin availability within the highly oxidative environment of the photoreceptor outer segment. Discussion Retinal metabolism must continuously meet extreme energetic and redox demands to sustain phototransduction, OS renewal, synaptic transmission, and the rapid membrane turnover characteristic of photoreceptors. These processes require tight coordination between mitochondrial energy production and cellular redox homeostasis. Riboflavin and its derivatives, FMN and FAD, are essential cofactors for numerous flavoproteins that drive oxidative metabolism, including complexes I and II of the mitochondrial electron transport chain, fatty acid β‐oxidation enzymes, and oxidoreductases involved in antioxidant defense [2, 3, 4, 10]. Through their ability to undergo reversible redox cycling, flavins support electron transfer reactions while also contributing to the regulation of cellular redox balance. Disruption of flavin‐dependent pathways can therefore impair mitochondrial respiration, increase reactive oxygen species production, and compromise antioxidant capacity, processes strongly linked to photoreceptor dysfunction and retinal degeneration [11, 13, 17, 31]. Within this metabolic framework, we previously identified RTBDN as a key extracellular regulator of retinal flavin homeostasis. RTBDN is expressed exclusively by rod photoreceptors and plays a central role in maintaining retinal riboflavin and total flavin pools [19, 39, 40, 44]. We demonstrated that RTBDN binds riboflavin both in vitro and in vivo [39, 40], and genetic ablation of RTBDN leads to early reductions in retinal flavin levels preceding overt degeneration [40]. Importantly, retinoid content remains unchanged in Rtbdn −/− mice, arguing against the hypothesis that RTBDN functions as an 11‐cis retinal transporter proposed by Fan et al. [55] and instead supporting its specific role in flavin regulation and retinal redox metabolism. RTBDN expression is not static but dynamically regulated by riboflavin availability and retinal stress [19, 34]. During dietary riboflavin deprivation, RTBDN levels initially increase, suggesting an adaptive response aimed at preserving extracellular riboflavin availability and sustaining flavin‐dependent redox reactions before declining with prolonged deficiency [19, 34]. In parallel, RTBDN expression is significantly elevated during the degenerative processes in several inherited retinal degeneration models [42, 43]. Because oxidative stress and mitochondrial dysfunction are hallmarks of photoreceptor degeneration, these findings suggest that RTBDN may participate in metabolic stress responses that attempt to stabilize flavin‐dependent redox pathways during disease progression. In the present study, we extend these observations by demonstrating that RTBDN is dynamically modulated across development, circadian light cycles, and nutrient availability. Developmentally, RTBDN expression begins between P5 and P7, coincident with early OS formation and prior to eye opening. Its levels increase progressively as OSs elongate, peak around P120, and decline by P360. Importantly, this temporal pattern parallels changes in total retinal flavin content. The onset of RTBDN expression coincides with the metabolic maturation of photoreceptors, during which differentiating cells transition toward high rates of aerobic glycolysis and oxidative phosphorylation to meet increasing energetic and biosynthetic demands [21, 27, 32]. These metabolic transitions require robust flavin availability to support electron transport and maintain mitochondrial redox balance. The peak of RTBDN expression during early adulthood corresponds to sustained OS renewal and synaptic activity, whereas its age‐associated decline mirrors the progressive reduction in mitochondrial efficiency and increased oxidative vulnerability observed in aging neural tissues [8]. These coordinated trajectories reinforce a functional link between RTBDN abundance and the maintenance of flavin‐dependent metabolic and redox capacity during photoreceptor maturation and aging. Light exposure further refines RTBDN regulation. Eye‐opening acts as a physiological trigger that increases both RTBDN levels and retinal flavin content, suggesting that sensory‐driven metabolic demand contributes to its upregulation. Light‐dependent regulation of metabolic pathways is well documented in the retina; photic stimulation alters mitochondrial activity, glycolytic flux, and antioxidant responses to accommodate fluctuating energy and redox requirements [27]. Retinal metabolism is also tightly coupled to circadian rhythms, with oscillations in oxidative metabolism, mitochondrial respiration, and flavin‐dependent enzymatic activity [46, 58]. Our findings indicate that RTBDN participates in this adaptive metabolic program during a developmental window when photoreceptor physiology becomes entrained to environmental light cycles [23, 24]. Such regulation may ensure that flavin availability remains synchronized with daily fluctuations in mitochondrial electron transport and ROS generation. A notable observation of this study is that RTBDN abundance remains relatively stable across the daily light–dark cycle, whereas its spatial distribution undergoes pronounced rearrangement. During the dark phase, RTBDN becomes enriched around cone photoreceptors, whereas after light onset, it redistributes toward the OS/RPE interface. Because RTBDN is expressed exclusively by rods [39], its preferential localization near cones during darkness suggests an interphotoreceptor metabolic support role. Photoreceptor metabolic demand is elevated in darkness due to the dark current and increased mitochondrial activity required to sustain ion pumping and phototransduction readiness [20, 32]. Cones, which exhibit high ATP and GTP turnover rates, may therefore require enhanced flavin‐dependent oxidative metabolism under dark‐adapted conditions [59] in preparation for light onset. By binding riboflavin, RTBDN may facilitate localized delivery of flavin cofactors to cones, thereby supporting mitochondrial electron transport and maintaining redox balance in cells with exceptionally high metabolic throughput. Nutrient availability further modulates RTBDN localization. Fasting reduced total retinal flavin levels and induced redistribution of RTBDN toward cone‐rich regions and the OS/IS interface without altering total RTBDN abundance. Refeeding alone did not restore RTBDN localization unless riboflavin was replenished, indicating that riboflavin status specifically regulates its spatial organization. These findings align with emerging concepts in redox biology whereby metabolic enzymes and cofactor‐binding proteins dynamically re‐localize in response to nutrient flux to optimize cellular energy production and redox buffering [60, 61, 62]. Similar adaptive redistribution has been reported for glycolytic enzymes and mitochondrial networks during metabolic stress [63, 64]. RTBDN movement between the OS/RPE interface, inner segments, and cone‐associated domains may therefore represent a mechanism for spatially prioritizing flavin availability in regions with the greatest metabolic and redox demand [27, 65, 66]. The age‐related decline in RTBDN and flavin levels may contribute to increased susceptibility to retinal degeneration. Reduced availability of FMN and FAD would be expected to impair flavoprotein‐mediated electron transfer reactions, decrease mitochondrial efficiency, and increase ROS production, ultimately compromising antioxidant defenses. Such mechanisms are strongly implicated in photoreceptor loss in multiple retinal degenerative diseases [67, 68]. Cones may be particularly vulnerable to these changes because of their high metabolic and oxidative burden. Although RTBDN variants have not yet been directly linked to cone‐specific human disease, RTBDN ablation results in rod‐cone dystrophy [40] and accelerates degeneration in disease models associated with PRPH2 and rhodopsin mutations [42, 43]. The preferential redistribution of RTBDN toward cones observed here, together with its decline during aging, raises the possibility that RTBDN may function as a metabolic modifier in diseases characterized by mitochondrial dysfunction and oxidative stress, including age‐related macular degeneration [68]. Taken together, our findings establish RTBDN as a developmentally regulated, riboflavin‐sensitive, and circadian‐responsive regulator of retinal flavin homeostasis. Rather than acting solely through changes in expression, RTBDN dynamically redistributes within the photoreceptor environment in response to developmental cues, nutrient status, and diurnal metabolic fluctuations. Through this spatial regulation of flavin availability, RTBDN may help sustain mitochondrial electron transport, limit ROS accumulation, and preserve redox balance in photoreceptors. These results expand our understanding of how cofactor metabolism integrates with retinal redox biology and identify RTBDN as a previously unrecognized component of the metabolic networks that maintain photoreceptor resilience. Author Contributions X.Z., M.R.A.‐U. and M.I.N. conceived and designed the research studies presented in this manuscript. X.Z. performed experiments, analyzed data, prepared the illustrations, and wrote the initial draft of the manuscript. M.S.M. collected and sectioned samples for the study. G.P.M. conducted High‐Performance Liquid Chromatography for retinoids. M.I.N. and M.R.A.‐U. revised and finalized the manuscript. Funding This work was supported by grants from the National Institutes of Health (EY10609 to MIN&MRA, EY034671 to M.I.N., and EY037470 to M.R.A.) and the Foundation Fighting Blindness (TA‐GT‐0624‐0889‐UTX) to M.R.A. Conflicts of Interest The authors declare no conflicts of interest. Supporting information Table S1: List of antibodies used in the current study. Figure S1: RTBDN and PRPH2 expression during retinal development. (A) Developmental immunohistochemical analysis of paraffin‐embedded mouse retinal sections at the indicated ages, labeled for RTBDN (red), PRPH2 (green), PNA (white), and nuclei counterstained with DAPI (blue). Both RTBDN and PRPH2 are first detected at P7. IS, inner segment; ONL, outer nuclear layer; OS, outer segment. Scale bar: 2 μm. Figure S2: Retinal flavin levels and retinal protein expression under dark‐reared (DR) and cyclic light (CL) conditions. (A, B) Retinal RTBDN (A) and PRPH2 (B) protein levels at P15, P21, and P30 were measured by western blot and normalized to actin in DR and CL groups across developmental time points. Data are presented as mean ± SD. Statistical significance was determined using the non‐parametric Mann–Whitney test. Sample size: N = 3 for all groups. p values: RTBDN/ACTIN: P15: p = 0.1000 (ns), P21: p = 0.7000 (ns), P30: p > 0.9999 (ns); PRPH2/ACTIN: P15: p = 0.1000 (ns), P21: p = 0.2000 (ns), P30: p > 0.3000 (ns). (C–E) Quantification of total flavin, RF, FMN, and FAD levels in retinas collected at P15 (C), P21 (D), and P30 (E) from mice maintained under DR or CL conditions. Data are presented as mean ± SD. Statistical significance was determined using the non‐parametric Mann–Whitney tests. Sample size: N = 3 for all samples. Figure S3: Levels of riboflavin, FMN, and FAD in mouse retinas under light deprivation and different dietary conditions. (A) Experimental design and timeline. P30 mice were maintained under a controlled 12 h/12 h light/dark cycle. Mice were fasted and then fed for 5 h with either riboflavin‐deficient chow (RDC) or regular chow (RC). Retinas were collected at 12:00 p.m. for analysis. Timeline was generated using BioRender.com (accessed on February 16, 2026). (B) Retinal levels of riboflavin, FMN, and FAD under different dietary conditions (Fasted; Fasted > Fed RDC for 5 h; Fasted > Fed RC for 5 h; Fed RDC; Fed RC). Data are presented as mean ± SD. Statistical significance was determined using the Kruskal–Wallis test with Dunn's multiple comparisons test. Sample size: N = 6 for each experimental group. Figure S4: RTBDN localization under different dietary conditions. Representative confocal images of P30 retinal cross‐sections from mice under different dietary conditions (Fasted; Fasted > Fed RDC for 5 h; Fasted > Fed RC for 5 h; Fed RDC; Fed RC) labeled for RTBDN (red), acetylated tubulin (AcTUB, green), PNA (white), and DAPI (blue). Two dimensional (2D) views illustrate RTBDN distribution across the photoreceptor layer under different dietary conditions, whereas corresponding three dimensional (3D) views highlight the spatial relationship of RTBDN with rod and cone photoreceptors. IS, inner segments; ONL, outer nuclear layer; OS, outer segments. Scale bar: 5 μm. Figure S5: Retinal flavin levels and RTBDN localization during different times of the day. (A) Quantification of total flavin, RF, FMN, and FAD levels in retinas collected at 6:00 a.m. and 3:00 p.m. Data are presented as mean ± SD. Statistical significance was determined using the non parametric Mann–Whitney tests. Sample size: N = 3 for all samples. (B) Representative confocal images of P30 retinal cross‐sections collected at 6:00 a.m., 8:00 a.m., 11:00 a.m., 1:00 p.m., and 3:00 p.m., labeled for RTBDN (red), acetylated tubulin (AcTUB) (green), PNA (white), and DAPI (blue). Two‐dimensional (2D) views illustrate RTBDN distribution across the photoreceptor layer at each time point, whereas corresponding three‐dimensional (3D) views highlight the spatial relationship of RTBDN with rod and cone photoreceptors (lower panels). IS, inner segments; ONL, outer nuclear layer; OS, outer segments. Scale bar: 5 μm. Figure S6: Full‐length immunoblots corresponding to cropped RTBDN western blot images shown in the main figures. Representative full, uncropped western blots for RTBDN immunoblotting are shown for Figures 1, 3, 4, 5, and 7. Molecular weight markers and all detected bands are included. Arrows indicate the RTBDN‐specific bands presented in the corresponding main figures. Rtbdn−/− retinal extract was included as negative controls. IB, immunoblot. Figure S7: Representative immunofluorescence images from independent biological replicates demonstrating RTBDN localization in the retina under different conditions. (A, B) Representative retinal sections from independent biological replicates (N = 1–3) at P15 and P30 under DR and CL conditions stained for RTBDN (A, red) with EZRIN (A, green) or AcTub/PNA (B, green/white), with DAPI (blue). (C) Representative retinal sections from independent biological replicates (N = 1–3) collected under fed and fasted conditions at P30 and stained for RTBDN (red), AcTub/PNA (green/white), and DAPI (blue). (D) Representative retinal sections from independent biological replicates (N = 1–3) collected at different circadian time points (6:00 a.m., 8:00 a.m., 11:00 a.m., 1:00 p.m., and 3:00 p.m.) and stained for RTBDN (red), PNA (white), and DAPI (blue). Images shown are representative of staining patterns consistently.