Light Changes Gene Activity Related to Energy Production and Heme Processing in Zebrafish

The wildtype WIK zebrafish strain (WIK, Danio rerio) was maintained at 28 °C in water circulation systems, under 14:10 light/dark conditions, and fed twice daily [42] (general license for fish maintenance and breeding: Az.: 35-9185.64/BH KIT IBCS-BIP Karlsruhe Institute of Technology (KIT)). Six-month-old fish were crossed to produce embryos for a subset of our experiments. Crossing was performed according to standard methods [42]. All husbandry and experimental procedures were performed following European Legislation for the Protection of Animals used for Scientific Purposes (Directive 2010/63/EU). Embryos were raised for four days in complete darkness, in 28 °C incubators in E3 medium, with the addition of 200 µM 1-phenyl 2-thiourea (PTU) from 24 hpf onwards. At the beginning of the fourth day (96 hpf), embryos were exposed to blue light for up to 6 h (up to 102 hpf) and sampled during this period of illumination. Therefore, the embryos were sacrificed before the free-feeding stage was reached (120 hpf), at which point ethical board approval is required in Germany.

Zebrafish (PAC-2) [20] and cavefish (EPA) embryo-derived cell lines [23] were cultured in Leibovitz L-15 medium (Gibco - Thermo Fisher (Waltham, MA, USA)) supplemented with 15% or 20% Fetal Bovine Serum (FBS, Gibco), respectively, 100 U/mL of penicillin, 100 µg/mL of streptomycin, and 50 µg/mL of gentamicin (Gibco) in an atmospheric CO₂, non-humidified incubator at 26 °C. For all experiments, cells were seeded in 96-well, 24-well, or 6-well plates at 3 × 10⁴ cells/well, 8 × 10⁴ cells/well, or 3 × 10⁵ cells/well, respectively. As a proxy for elevated ROS, PAC-2 and EPA cells were treated with 300 µM H₂O₂, and in order to eliminate any possible direct effects of light exposure, all cells were maintained in constant darkness during the treatment period. Precise experimental details are described in the Figure Legends and Section 3.

Cell and embryo illumination was performed at a constant temperature using a monochromatic blue LED light source (λpeak = 468 nm, Kopa (Barcelona, Spain)) adjusted to deliver the photon flux (1.42 ± 0.04 × 10¹⁸ photons/s/m²). Precise details of lighting conditions in our experimental analysis are provided in the Figure Legends and Results sections.

After two days in darkness, PAC-2 and EPA cells were exposed to up to 6 h of blue light, and samples were collected in Trizol reagent (Thermo Fisher (Waltham, MA USA)) for RNA extraction. Total RNA was extracted according to the manufacturer's instructions. Total RNA (4 µg) from PAC-2 and EPA cells was shipped to Novogene (UK) for mRNA sequencing and preliminary bioinformatic analysis. For gene expression analysis, reverse transcription was performed with the RevertAid reverse transcription kit (Thermo Fisher) using random primers. SyBrGreen (Promega (Walldorf, Germany)) master mix was used for quantitative PCR analysis in the Real-Time PCR ABI QuantStudio3 qPCR Cycler (Thermo Fisher). Primers are shown in Supplementary Table S1. Data were analyzed according to the 2^−ΔΔCT method, and expression of β-actin was used for normalization.

Sequencing using the Illumina platform, as well as subsequent quality control (error rate distribution, GC-content distribution, and data filtering), mapping to the reference genome (GRCz11 for the D. rerio dataset), and gene expression quantification were performed by Novogene UK. Since there is no available reference genome for the cavefish P. andruzzii, de novo transcriptome analysis was performed for both species. Briefly, Trinity software version 2.13.2 [43] was used to assemble the short reads into full transcripts, which were then quantified and annotated via alignment of the transcripts and the longest Open Reading Frame (ORF) against UniProt via Blast-X version 2.13 alignment. Domains were predicted via alignment in the PFAM database. Transcripts within the datasets were quantified and excluded based on the Transcript Per Million (TPM) cut-off of 5. For all analyses, transcripts were quantified for each sample, and Differential Expression analysis was performed across all timepoints, considering transcripts with |logFoldChange| >= 1 and adjusted p-value < 0.001 to be differentially expressed. Finally, functional enrichment testing was performed via Gene Ontology (GO) analysis on 0 h vs. 6 h Differentially Expressed Genes (DEGs). A detailed pipeline for the analyses is shown in Figure 1.

Actinomycin-D was used as an inhibitor of cellular transcription. PAC-2 cells were treated with 5 µg/mL Actinomycin-D (A1410, Sigma-Aldrich (St. Louis, MO, USA)) before exposure to blue light (t = 0 h), and samples were collected in Trizol at 1, 3, and 6 h. Controls treated with Actinomycin-D were kept in darkness. Following RNA extraction and cDNA synthesis, RT-qPCR was used to measure the relative amounts of mRNA present in the samples. Primers used are listed in. Supplementary Table S1

Promoter region sequences, 1 kb upstream of the Transcription Start Site (TSS) plus the 5′ untranslated regions of genes, were retrieved using BioMart from Ensembl release 109 (GRCz11). Clover [44] was used to screen the promoters for putative D-box and E-box sequences. Putative light-responsive promoter fragments containing D-box-like and E-box-like motifs were cloned into the pGL3Basic (Promega) luciferase expression vector. A smaller fragment of hebp2 (144 bp) was cloned into the minimal promoter luciferase expression vector pLucMCS (Stratagene (La Jolla, CA, USA)). The sequences of single, predicted D-box and E-box sequences were mutated (4–6 bp each) using the Q5^® Site-Directed Mutagenesis kit (Promega) with the aim of disrupting the enhancer function.

Rapid Amplification of cDNA Ends (RACE) PCR (SMARTer RACE 5′/3′ Kit, Takara Bio (San Jose, CA, USA)) was used to amplify the 5′ and 3′ ends of the HLF-1, TEF-2, and Nfil3-1b sequences of the Somalian cavefish according to the manufacturer's instructions. The sequences for the other PAR-bZip and Nfil3 transcription factors were obtained from a partial genome sequencing dataset after confirmation via alignment with the zebrafish counterparts. The complete mRNA coding sequences were amplified from EPA cell cDNA and cloned into the CMV-driven pCS2-MTK vector for expression in eukaryotic cells. The start codon was mutated so that a 5′ 5x myc-tag sequence (EQKLISEEDL) was incorporated adjacent to the N-terminus of each PAR factor. All cloning was confirmed first by sequencing (Microsynth Seqlab, Göttingen, Germany) and then by their expression in transiently transfected cells via Western blot analysis for the myc-tagged proteins (see Supplementary Figure S2). The zebrafish PAR-bZip and Nfil3 factors studied were cloned previously [34].

PAC-2 cells in 6-well plates were transfected with 1 µg of the zebrafish and cavefish PAR-bZip expression vectors and kept in darkness for 48 h before lysis with Passive Lysis Buffer (Promega). Samples were run on 10% polyacrylamide-SDS gels together with the Color Prestained Protein Standard, Broad Range (10–250 kDa, New England BioLabs (Ipswich, MA, USA)), and blotted onto Immobilon^®-P PVDF membrane (Merck Millipore, Tullagreen, Ireland). A mouse anti-Myc tag antibody (1:1000, Sigma-Aldrich) and goat anti-mouse polyclonal antibody (1:7500, Sigma) were used to detect protein expression. Visualization was performed by using the ECL detection system (Bio-Rad, (Hercules, CA, USA)). Images were acquired and analyzed by using Image Lab Software version 6.1.0 build 7 (Bio-Rad). Western blot data are presented in Supplementary Figure S2.

Transfection was performed 24 h following cell seeding using FuGene HD (Promega), as described elsewhere [4,23,24]. For in vivo bioluminescence assays, 25–100 ng of the luciferase reporter plasmid was transfected into each well of a 96-well plate. After 24 h, 0.5 mM D-luciferin Firefly potassium salt (FL08607, Biosynth, (Bratislava, Slovakia)) was added to the culture medium, and bioluminescence was measured each hour using Topcount NXT or Envision counters (Perkin Elmer (Shelton, CT, USA)). For in vitro experiments, cells were transfected with 50–200 ng of the luciferase reporter vector, 50 ng of the β-galactosidase expression vector (pcDNA3.1/myc-His/lacZ, Invitrogen (Carlsbad, CA, USA)), and 1 ng of the pCS2-MTK expression vector containing the transcription factor of interest and lysed after 48 h using Firefly Lysis Buffer. The cell lysates were used for luciferase assays, and a β-galactosidase assay served as a control for transfection efficiency normalization. All transiently transfected fish cell culture experiments and standard molecular cloning methodology were performed according to local regulations for biology safety level 1 at Karlsruhe Institute of Technology (Permission number KITGF.KA.10.02)).

Data were analyzed and graphs were produced using R software version 4.3.2 and InkScape version 1.4.2. All results are expressed as means ± Standard Error of the Mean (SEM). ANOVA with multiple comparisons post hoc tests (Tukey HSD) or Bonferroni corrections were used to determine significance. Values of p < 0.05 were considered to be statistically significant, and values of p < 0.05, p < 0.001, and p < 0.001 are represented in graphs by *, **, and ***, respectively. Detailed statistical information and the data used to make the figures are reported in Supplementary Dataset S1.

In order to obtain a view of the full scope and, ultimately, the functional significance of light-induced gene expression in zebrafish cells, we initially characterized the light-induced transcriptome in the PAC-2 zebrafish cell line (originally established from zebrafish embryos [20]). Thus, PAC-2 cells were first kept in complete darkness for two days to ensure significant downregulation of light-induced gene expression as well as a dampening of clock-regulated gene expression, and then they were exposed to one, three, or six hours of Blue Light (BL) prior to RNA extraction. A monochromatic blue light source (λpeak = 468 nm) was chosen based on previous studies demonstrating a robust transcriptional response upon exposure to this wavelength [23]. Furthermore, by sampling at these timepoints, elevated mRNA levels were detected for all previously characterized light-regulated genes [4,23,45]. A control sample not exposed to blue light was harvested as time 0 (dark control, DD) (see the experimental scheme in Figure 2A). RNA extracted from the cell samples was subjected to quality control assurance (Nanodrop and Agilent 5400 assay systems) before sequencing and performing bioinformatic analysis. The resulting dataset can be downloaded from GEO (GSE290582).

Differentially Expressed Genes (DEGs) in zebrafish cells were identified as having a |log₂FoldChange| > 1 and p-value adjusted for multiple comparisons p-adj < 0.001. A total of 331 genes were differentially expressed after either 1, 3, or 6 h of blue light exposure compared to the DD control. Blue light mostly led to the upregulation of genes at 3 and 6 h (Figure 2B–D for up- and downregulated genes, respectively). Gene Ontology (GO) analysis of the genes upregulated at 6 h was performed to categorize the biological processes, cellular compartments, and molecular functions affected by light exposure. The top 25 enriched GO terms are shown in Figure 2E. As expected, the gene set was significantly enriched in the GO terms related to circadian clock regulation and circadian clock entrainment by photoperiod. Other significantly enriched categories included DNA damage response and DNA repair, as well as response to hydrogen peroxide and oxidative stress, detoxification processes, and glutathione biosynthesis. Interestingly, the GO results also revealed significant enrichment of processes related to mitochondria structure and function, as well as to heme biosynthesis and metabolism.

Among the new classes of genes found to be light-regulated were genes involved in the biosynthesis and transport of heme and other porphyrins (abcb6a, blvra, fech, slc40a1/ferroportin1, slc48a1a, tfr1a, and tspo), as well as genes encoding heme-binding proteins and hemoproteins (txbas1, cyb5a, cyp1ae1, hebp2, and soul5). Among the upregulated mitochondria-related genes were those coding for subunits of the ETC complexes and their assembly (sdha, sdhb, sdhaf1, sdhaf2, sdhaf3, afg3l2, chchd4b, and ttc19), as well as genes involved in the regulation of mitochondrial architecture and connectivity (mff, adck1, rap1gd1, retsat, and trf1a), and those implicated in the regulation of mitochondrial activity and membrane permeability (hebp2, bbc3, retsat, and sirt4). A total of 63 significantly upregulated genes belonging to these categories were identified as a group, which will hereafter be referred to as mitochondria and heme-related genes. A detailed list of all genes belonging to this group can be found in Supplementary Dataset S1. To validate our RNA-seq results, a subset of genes was selected based on the strength of induction and their established relevance [46,47,48,49,50]. Their induced expression, together with that of light-regulated controls, clock genes (per2 and cry1a), and a DNA-repair gene (6-4phr), was confirmed via RT-qPCR. Figure 2F represents the RT-qPCR results compared with data obtained by RNA-seq analysis. These results are also consistent with no significant changes in the expression of light-responsive genes occurring in the cell cultures under DD conditions during the 6 h sampling period for the RNA-seq and RT-qPCR analysis (see Figure S1).

In the Somalian cavefish (P. andruzzii), we previously reported that light/ROS-induced expression of clock genes and DNA-repair genes is absent [4,23,24,26]. Does this pattern extend to all other light-induced transcripts identified in the zebrafish cell line? Providing answers to this question allows us to assess whether the D-box control mechanism serves as a global regulator of light and ROS-induced gene expression. Thus, as the next step, we compared the light-induced zebrafish transcriptome with RNA-seq data from EPA cells (a cell line previously established from P. andruzzii embryos [23]) using the same experimental design.

To date, a complete annotated genome sequence is not available for P. andruzzii. Therefore, the transcriptome sequences of both the cavefish and zebrafish cell lines were initially assembled with a de novo pipeline before comparison and analysis (Figure 1). DEGs were identified with the same thresholds as the zebrafish dataset. The analysis showed that the magnitude of transcriptional changes occurring in response to blue light in cavefish cells was much smaller than in zebrafish cells. In the PAC-2 cells, a total of 607 DEGs were identified at 1, 3, or 6 h of exposure to blue light compared to the DD control, while only 278 DEGs overall were identified in cavefish. For both species, the number of identified DEGs increased with increasing exposure to blue light, with few genes being up- or downregulated after 1 h of exposure (Figure 3A,B). Furthermore, zebrafish cells displayed a more sustained response to light, with 154 genes (33% of all upregulated genes) consistently upregulated after both 3 and 6 h of exposure (Figure 3C). In contrast, in the transcriptomic response to light of cavefish cells, only 30 genes were upregulated (16% of all upregulated genes) at both 3 and 6 h, suggesting a more transient pattern of gene upregulation (Figure 3D).

While the GO analysis indicated significant enrichment of mitochondrial and heme-related processes in the zebrafish dataset, no such enrichment was observed in the cavefish dataset (Supplementary Dataset S1). GO terms related to circadian clock entrainment, response to light, and DNA repair were also not significantly enriched in cavefish. The expression of selected genes from the mitochondrial and heme-related gene pool was confirmed with RT-qPCR for both species (Figure 3E,F). Consistently reflecting the RNA-seq findings, all genes except abcb6a were significantly upregulated in zebrafish but not cavefish cells. abcb6a was upregulated in both cell types according to both RNA-seq and RT-qPCR analyses.

Visible light has been shown to induce ROS production in zebrafish cells, and the D-box enhancer is known to be necessary for the response to ROS, providing one mechanism by which light activates D-box-driven transcription. Previous studies have demonstrated that ROS increases alone significantly upregulate the expression of the D-box-controlled circadian clock and DNA damage repair genes in PAC-2 cells [23,24,45]. To test whether this also applies to abcb6a, hbp2, and soul5, PAC-2 cells were treated with 300 μM hydrogen peroxide, and gene expression was assessed via RT-qPCR compared with controls kept in constant darkness (Figure 4A). The treatment led to the upregulation of all three genes after 3, 6, and 9 h of treatment (Figure 4B).

Our transcriptomic and RT-qPCR gene expression analyses revealed that blue light upregulates the expression of various functional classes of genes in zebrafish cells, but not in cavefish cells. The light-induced expression pattern of the mitochondria and heme-related genes closely resembles that of light-responsive circadian clock and DNA-repair genes, whose expression has previously been documented in vivo in the context of the zebrafish larva. Thus, we examined whether this is also the case for this class of mitochondria and heme genes. We illuminated 4 dpf zebrafish embryos with blue light for up to 6 h and collected samples to perform RT-qPCR (Figure 4C). Our choice of this zebrafish developmental stage was based on the documented emergence of a functional circadian clock and light-regulated clock gene expression by 4 dpf [51]. The blue light exposure led to a significant increase in the expression of the abcb6a, hebp2, and soul5 genes, as observed for the clock gene per2 used as a positive control (Figure 4D). This pattern in zebrafish embryos mirrored the results seen in PAC-2 cells and is consistent with previous reports for circadian clock and DNA repair genes [36].

Is this increase in RNA expression in response to light purely due to transcriptional regulation, as for the clock and DNA repair genes? To address this question, we initially tested whether the mRNA stability of the mitochondrial and heme-related genes changed under blue light exposure. PAC-2 cells were treated with Actinomycin-D (5 µg/mL), an inhibitor of de novo gene transcription, immediately before exposure to 1, 3, or 6 h of blue light or darkness as a control. This experimental approach blocks any transcription activation and tests for any changes in mRNA stability (Figure 4E). c-myc mRNA levels were monitored as a positive control since this transcript is well documented to have a high turnover [52]. This treatment resulted in no significant difference in degradation time for all the mRNAs analyzed between the samples maintained in DD and the samples exposed to blue light. This indicates that the increased expression of these genes is the result of light-mediated transcriptional activation (Figure 4F).

It has previously been shown that the expression of many light-induced genes also oscillates in a circadian manner due to the presence in their promoters of canonical or non-canonical circadian enhancer elements termed E-boxes [27]. Therefore, rhythmic expression of these genes persists for several cycles when transferred to constant darkness. We next tested whether this was also the case for the mitochondrial and heme-related genes. PAC-2 cells were entrained to 12 h:12 h light:dark cycles (LD) for five consecutive days, followed by transfer to constant darkness or LD conditions once again, and gene expression was measured at regular intervals over a 28 h period (Figure 4G). The expression of the per1b gene, which exhibits similar oscillations in both LD and DD conditions, was used as a positive control for a purely circadian clock-regulated gene. Instead, the expression of the per2 gene, which exhibits oscillation only under LD conditions, was used as a positive control for a purely light-responsive gene. Our results showed that abcb6a, hebp2, and soul5 all behave as exclusively light- and not circadian clock-regulated genes (Figure 4H).

Which mechanism drives the light-dependent activation of abcb6a, hebp2, and soul5 gene expression? The findings above resemble the response to light of per2 and 6-4phr expression, which is the result of light-mediated transcriptional activation via D-box enhancer elements [4,27]. Therefore, as a first step, we performed a bioinformatic promoter analysis with Clover software (release date 2012-02-16) [44] to test for the presence of D-box enhancers in the promoters of our 63 genes of interest. The software predicted putative D-box sequences within the 5′ Untranslated Region (UTR) and 1 kb upstream of the Transcription Start Site (TSS) in 70% of mitochondria and heme-related genes (the results are shown in Supplementary Dataset S1). In addition, putative E-box sequences were predicted in 48% of these genes. In the case of our three selected genes, three putative D-box and one E-Box sequences were identified within the promoters of hebp2 (Figure 5A, above) and soul5 (Figure 5B, above), while three D-box and two E-box sequences were identified in the abcb6a promoter (Figure 5C, above). Indeed, the presence of E-box-like sequences in the proximity of D-boxes is consistent with cooperation between these two enhancer elements to mediate light-induced expression, as we previously reported for the regulation of per2 gene expression [27,45].

To directly test the contribution of these enhancer elements to directing light-regulated transcription of these genes, fragments of each promoter, encompassing the putative light-responsive promoter sequences, were cloned into a luciferase reporter expression vector, hereafter referred to as hebp2-Luc, abcb6a-Luc, and soul5-Luc. The constructs were then transfected into light-responsive PAC-2 zebrafish cells, as well as into non-light-responsive cavefish EPA cells, and then monitored for their luciferase expression during two days under LD cycles followed by transfer to two days in complete darkness (DD) (Figure 5A–C, middle). All three promoters showed robust light-induced bioluminescence during LD in zebrafish PAC-2 cells that was not maintained under DD conditions, indicating that our promoter constructs included the regulatory sequences required for the light responsive transcription observed with the endogenous genes. Instead, transfection of all three bioluminescent reporters in EPA cells did not elicit a rhythm of bioluminescence in either LD or DD conditions. In addition, consistent with our previously published results regarding D-box-regulated expression of the 6-4phr gene, the three bioluminescence promoter reporters transfected in zebrafish cells were also induced following H₂O₂ treatment (Figure 5A–C, below). This again mirrors endogenous mRNA expression of these genes (Figure 4B), as observed for the promoters of the cry1a and per2 genes [23,45].

To further investigate the specific role of each D-box and E-box sequence in light-mediated gene expression, we systematically mutated each D-box and E-box in the hebp2-Luc reporter via site-directed mutagenesis, resulting in four new constructs where only a single D-box or E-box was retained. In addition, a control construct where all D-box and E-box sequences were mutated simultaneously was also generated (Figure 6A).

When transfected into PAC-2 cells and exposed to an LD cycle, all the mutated constructs in an in vivo luciferase assay showed attenuated activation in response to light compared to the intact, non-mutated hebp2-Luc reporter (Figure 6B). Treatment with H₂O₂ did not activate the mutated constructs (Figure 6D–I). An in vitro luciferase assay of cells transfected with these same constructs and exposed to blue light for eight hours also revealed significantly reduced activation of the mutated promoters measured at a single timepoint (Figure 6C). These assays indicate that the induction of hebp2-Luc by light relies on the function of both D-box and E-box sequences and that a synergistic effect between these enhancers underlies the light- and ROS-dependent induction of hebp2.

Previous reports have demonstrated that PAR-bZip transcription factors bind to the D-box enhancer and regulate the transcription of various light-induced genes, including circadian clock and DNA repair genes [24,26,45,53]. To test whether this is also the case for the hebp2 promoter, we compared the activation mediated by each zebrafish PAR-bZip homolog [34] in an in vitro luciferase assay using the hebp2-Luc reporter in PAC-2 cells (Figure 7A,B). Our results reveal a range of activation of reporter gene expression, with the strongest activation observed for the TEF and HLF factors. Cavefish-specific loss of PAR-bZip factor function has been proposed as a potential contributor to the loss of light-induced gene expression in P. andruzzii. Thus, we amplified the coding sequences of all these P. andruzzii factors by RT-qPCR from EPA cell RNA and cloned them into the pCS2-MTK expression vector. The predicted P. andruzzii and zebrafish amino acid sequences were aligned, revealing that the C-terminal portions of the PAR-bZip proteins, containing the PAR, basic, and leucine zipper domains, are highly conserved between these species, while the N-terminal sequence of each factor is more divergent (Table 1). Despite the absence of premature truncation mutations that we previously reported in various opsin, clock, and DNA repair genes [4,26], the cavefish factors exhibited non-conservative amino acid substitutions in their N-terminal regions, which could potentially affect their structure and function. Might the amino acid differences observed between these two cyprinid species account for the significant deficiency in light-induced hebp2 gene expression? To address this question, we initially compared the activation of hebp2 expression mediated by the zebrafish and P. andruzzii homologs of TEF and HLF in transfected PAC-2 cells (Figure 7C). Ectopic expression of the recombinant PAR-bZip factors was initially confirmed by Western blotting analysis of the myc-tagged zebrafish and cavefish proteins (Figure S2).

Our results showed that both the zebrafish and the cavefish factors robustly activated hebp2-Luc reporter expression in PAC-2 cells, with some species-specific differences in the observed amplitude of reporter gene activation (Figure 7C). However, repeating these transfection assays in cavefish EPA cells resulted in considerably lower levels of reporter activation for both cavefish and zebrafish factors (Figure 7D). These results support the notion that the amino acid sequence differences in the cavefish PAR-bZip factors do not account for the loss of light and ROS-induced D-box mediated hebp2 transcription in P. andruzzii. Instead, they implicate the loss of function of distinct regulatory cofactors in cavefish, which thereby attenuate PAR-bZip transcription factor-mediated activation in response to light and ROS.

Our transcriptomic analysis has revealed significant upregulation of genes associated with mitochondrial structure and function, potentially leading to alterations in permeability, ATP output, and overall mitochondrial function with general consequences for energy metabolism. Considering the key importance of mitochondria and electron transport chain enzymes in mediating blue light-induced apoptosis [18,19], these factors could be part of an adaptive and protective mechanism counteracting the challenges posed by sunlight exposure, preserving mitochondrial integrity, and promoting cellular energy production.

The upregulation of heme-related genes further underscores the multifaceted cellular response to light. Heme is synthesised predominantly in the mitochondria and serves as a crucial cofactor for numerous enzymes within and outside this organelle [54,55,56]. It plays a pivotal role in managing oxidative stress, as a prosthetic group for various antioxidant enzymes that break down ROS molecules [2,57]. Heme is also involved in xenobiotic metabolism by serving as a cofactor for cytochrome P450 enzymes, which oxidize compounds to facilitate their excretion [58,59]. In addition, mouse knockouts of the three PAR-bZip factors have revealed a role in mediating the expression of genes involved in the metabolism of xenobiotics [39]. Furthermore, circadian rhythms of drug metabolism have been extensively documented [60,61]. Therefore, several lines of evidence support an important role for sunlight exposure in heme homeostasis both via transcriptional regulation of crucial enzymes and via affecting heme transport and metabolism. Given the loss of light and ROS-induced expression of mitochondrial and heme-related genes in P. andruzzii cells, it is tempting to speculate that this cavefish may consequently exhibit increased sensitivity to acute sunlight exposure, with a potentially corresponding increase in cell death. Alternatively, these regulatory differences might reflect an adaptive strategy to survive the extreme, perpetually dark cave environment. Clearly, additional studies will be required to evaluate the functional significance of these cavefish-specific changes in gene regulation networks.

In this study, we identified functional D-box and E-box sequences in the promoters of 70% zebrafish light-regulated genes related to mitochondrial function and heme metabolism. Promoter fragments of hebp2, abcb6a, and soul5 containing these enhancers were shown to drive light-inducible transcriptional activation. The E-box enhancer serves as the binding site for the Clock-Bmal heterodimer and thereby typically acts as a regulatory target for the core machinery of the circadian clock [62]. It is therefore surprising that none of the genes of this functional class showed evidence for circadian rhythmicity of expression under constant darkness conditions. However, other members of the bHLH PAS domain family of transcription factors may also bind to E-box-like enhancers [63]. Therefore, potentially, non-clock-regulated factors may occupy these sites and enable regulation via the D-box. The D-box-binding sequence was initially described as the canonical sequence 5′–RTTAYGTAAY–3′ (R is A or G, and Y is C or T) [30], but subsequent research has broadened the scope of D-box sequences. In the present study, one of the identified D-box sequences of the hebp2 promoter is canonical (5′–GTGATGTAAC–3′), while the other two represent variants of this sequence (5′–GTTACTTAAC–3′ and 5′–GTTATTTAAG–3′). Interestingly, the PAR-bZip factors exhibited differential preference for hebp2 and cry1a light-responsive regions. The strongest activators for hebp2 were the TEF and HLF factors, while HLFs and DBP-2 preferentially activated cry1a [45]. This variability potentially highlights promoter-specific differences in the regulatory role of D-boxes. In addition, heterodimerization, post-translational modification, and other regulatory mechanisms may influence the function of these transcription factors. Further studies, such as studying knockout models for individual PAR-bZip factors, will be required to dissect the specific roles of these factors in D-box-mediated transcription in response to sunlight and ROS.

Comparing the transcriptional response to light in zebrafish with that in the Somalian blind cavefish P. andruzzii is a valuable comparative system for elucidating the underlying light- and ROS-regulated mechanisms, as well as understanding how they have been shaped by evolution in constant darkness [4,23,24,26]. In the present study, we reveal that zebrafish show a more robust and sustained transcriptional response to light compared to cavefish. Only a few cavefish genes have retained light-inducible expression, notably abcb6a and ddb2. We have previously demonstrated the necessity of an E2F binding site cooperating with a D-box enhancer for the induction of ddb2 gene expression in response to visible light, UV, and ROS exposure in cavefish cells, and it is tempting to speculate that a similar mechanism operates for abcb6a [24]. Given the retention of functional D-box sequences in the promoters of clock and DNA repair genes in P. andruzzii, it is tempting to speculate that altered functionality of the cavefish PAR-bZip factors contributes to this adaptation. However, the C-terminal truncation mutations that we previously documented in a subset of opsin, core clock genes, and DNA repair genes from P. andruzzii [4,26] are absent from the PAR-bZip factor genes. Furthermore, the cavefish TEF and HLF homologs robustly activate transcription from the hebp2 promoter in zebrafish cells, but not in cavefish cells. These results point to the presence of additional light and ROS-regulated cofactors, acting upstream of the PAR-bZip transcription factors, which exhibit an attenuated function in P. andruzzii. It is also conceivable that P. andruzzii carries loss-of-function mutations affecting ROS-responsive epigenetic factors, which may modulate D-box enhancer function by altering chromatin state, DNA methylation, or transcription factor accessibility, thereby controlling gene expression in a light-dependent manner [64]. During vertebrate evolution, major changes have occurred in the regulation and function of D-box binding transcription factors. Specifically, while in mammals, D-box-mediated transcriptional regulation operates in the context of clock output pathways, in fish, D-boxes serve as targets for light input pathways for the clock. Therefore, from a more general perspective, these results point to remarkable plasticity in the function of D-box-regulated transcription over the course of vertebrate evolution.