Expression and localization of the aryl hydrocarbon receptors and cytochrome P450 1A during early development of Atlantic cod (Gadus morhua)

The aryl hydrocarbon receptor (Ahr) is a ligand-activated transcription factor that mediates the toxicity of di- oxins and dioxin-like compounds (DLCs) in vertebrates. Two clades of the Ahr family exist in teleosts (Ahr1 and Ahr2), and it has been demonstrated that Ahr2 is the main protein involved in mediating the toxicity of dioxins and DLCs in most teleost species. Recently, we characterized the Atlantic cod ( Gadus morhua ) Ahr1a and Ahr2a receptors. To further explore a possible subfunction partitioning of Ahr1a and Ahr2a in Atlantic cod we have mapped the expression and localization of ahr1a and ahr2a in early developmental stages. Atlantic cod embryos were continuously exposed in a passive-dosing exposure system to the Ahr agonist, benzo[ a ]pyrene (B[a]P), from ﬁ ve days post fertilization (dpf) until three days post hatching (dph). Expression of ahr1a , ahr2a , and the Ahr-target genes, cyp1a and ahrrb, was assessed in embryos (8 dpf and 10 dpf) and larvae (3 dph) with quan- titative real-time PCR analyses (qPCR), while in situ hybridization was used to assess the localization of expression of ahr1a , ahr2a and cyp1a . Quantitative measurements showed an increased cyp1a expression in B[a]P- exposed samples at all sampling points, and for ahr2a at 10 dpf, con ﬁ rming the activation of the Ahr-signalling pathway. Furthermore, B[a]P strongly induced ahr2a and cyp1a expression in the cardiovascular system and skin, respectively, of embryos and larvae. Induced expression of both ahr2a and cyp1a was also revealed in the liver of B[a]P-exposed larvae. Our results suggest that Ahr2a is the major subtype involved in mediating re- sponses to B[a]P in early developmental stages of Atlantic cod, which involves transcriptional regulation of biotransformation genes, such as cyp1a . The focused expression of ahr1a in the eye of embryos and larvae, and the presence of ahr2a transcripts in the jaws and ﬁ n nodes, further indicate evolved specialized roles of the two Ahrs in ontogenesis.


Introduction
The aryl hydrocarbon receptor (Ahr) is a ligand activated transcription factor that has existed for more than 600 millon years of animal evolution, and it is broadly distributed in multiple phyla . Ahr is a member of the basic helix-loop-helix/Per-Arnt-Sim family and has been extensively studied in toxicology because of its role in mediating responses to many xenobiotic compounds in vertebrates (Nebert, 2017;Okey, 2007). As a result of a tandem duplication of Ahr prior to the divergence of the cartilaginous and bony fish T 1974; Barron et al., 2004;Goksøyr et al., 1994;Hylland et al., 2008;Sørhus et al., 2015). Ahr2 is the main protein involved in mediating xenobiotic responses to such compounds in most fish species (Tanguay et al., 1999;Walker et al., 1991;Zanette et al., 2009), as demonstrated in loss-of-function studies using morpholino-modified antisense oligonucleotides or genome editing (Clark et al., 2010;Garcia et al., 2018;Goodale et al., 2012;Jönsson et al., 2007Jönsson et al., , 2009Prasch et al., 2003;Sugden et al., 2017).
Unliganded Ahr is located in the cytoplasm in a protein complex with two HSP90 proteins, a co-chaperon protein p23, and Ahr-interacting protein (AIP) (Carver and Bradfield, 1997;Denison et al., 2011;Kazlauskas et al., 1999;Perdew, 1988). The recognition and binding to an agonist promote the conformational activation of Ahr and subsequent translocation into the nucleus where Ahr heterodimerizes with the aryl hydrocarbon receptor nuclear translocator (Arnt). This heterodimer binds to xenobiotic response elements (XRE) upstream of Ahr target genes and modulate the transcription of different enzymes involved in the biotransformation of xenobiotics, including cytochrome P450 1A (Cyp1a) (Denison et al., 2011;Pollenz et al., 1993;Soshilov and Denison, 2008;Swanson, 2002;Whitlock, 1999). Another Ahr target gene is the aryl hydrocarbon receptor repressor (Ahrr), which has a role in regulating Ahr transcriptional activity. Different mechanisms of Ahrr-mediated repression of the activity of Ahr have been postulated, including a crystallography study that showed the formation of an Ahrr-Arnt heterodimer (Mimura et al., 1999;Sakurai et al., 2017).
During the last decades different studies have highlighted the role of Ahr in signalling pathways and physiological systems other than the response to xenobiotic compounds, including the cardiovascular, reproductive and immune system in vertebrates (Esser and Rannug, 2015;Mulero-Navarro and Fernandez-Salguero, 2016;Nebert, 2017). It has been suggested that the original function of Ahr was to regulate the development of neural systems and sensory structures, and that these ancestral functions are still present in modern living invertebrates . Hahn et al. suggested a specialization of the different ahr paralogs through subfunction partitioning in some nonmammalian vertebrates . Some of the partitioning mechanisms described include tissue-specific expression patterns, as well as ligand-and target gene specificity. An example is the zebrafish, whose three Ahrs are thought to have evolved different functional roles. While the Ahr2 of zebrafish (zfAhr2) is involved in mediating toxic responses to dioxin-like compounds and PAHs, zfAhr1b is suggested to have a role in embryonic development (Karchner et al., 2005). zfAhr1a, on the other hand, is not activated by any classical Ahr agonists, but may still be involved in the regulation of physiological responses by being activated by a variety of non-halogenated compounds (Chlebowski et al., 2017;Goodale et al., 2012;Incardona et al., 2006;Knecht et al., 2013). A role of zfAhr2 in the development of the nervous and sensory system has also been demonstrated (Garcia et al., 2018;Goodale et al., 2012;Souder and Gorelick, 2019).
Atlantic cod (Gadus morhua) has for many years been used as an indicator species in marine pollution monitoring programs and field studies (Beyer et al., 1996;Dale et al., 2019;Goksøyr et al., 1994;Hylland et al., 2008;OSPAR, 2010;Sundt et al., 2012). However, in order to obtain a better understanding of how cod mediates responses to pollutants, we recently functionally described the Ahr1a and Ahr2a receptors for the first time (Aranguren-Abadía et al., 2020). These paralogous genes differed in their tissue-specific expression in juvenile cod, as well as possessing different ligand binding affinities and transactivation activities, which may indicate a subfunctional partitioning of the Ahr receptors in this species.
The aim of the present study was to further explore a possible subfunction partitioning of Ahr1a and Ahr2a in Atlantic cod by mapping the expression and localization of ahr1a, ahr2a and cyp1a in early developmental stages. Activation of the Ahr-signalling pathway was confirmed by measuring expression of cyp1a in B[a]P-exposed samples. Our results suggest that Ahr2a is the major subtype involved in mediating responses to B[a]P in early developmental stages of Atlantic cod. Furthermore, expression of ahr2a in developing jaws and fin nodes of larvae, and the persistent presence of ahr1a transcripts in the eye of both cod embryo and larvae indicates a physiological role of these genes in the ontogenesis of these organs.

Material and methods
2.1. Passive-dosing exposure experiment 2.1.1. Cleaning and loading of silicone rods Water exposure of highly hydrophobic compounds like B[a]P is challenging due to the very low water solubility and high absorption to surfaces in the exposure tanks. Passive dosing was chosen to establish and maintain freely dissolved concentrations of B[a]P at the solubility limit, while avoiding co-solvent addition (Smith et al., 2010). Translucent 1-mm silicone rods (Altec Extrusions, Victoria, UK) were cut to 1 m lengths and precleaned by soaking in methanol and ultrapure water washes. All chemicals used were GC Suprasolv® analytical grade and supplied by Merck (Darmstadt, Germany). The rods were then loaded into saturated B[a]P methanol solutions for two days, whereas controlrods were immersed in methanol (> 99 % purity, Sigma-Aldrich, St. Louis, Missouri, United States) (Smith et al., 2010). After the loading, excess methanol was removed by three successive soaks in ultrapure water for 2-24 hours. One corresponding rod was allowed to equilibrate with 80 mL of autoclaved seawater in each glass beaker for two days before the start of the exposures. All glassware was heated to 400°C overnight before use.

Exposure of Atlantic cod eggs
Eggs and sperm were stripped from brood stocks of Atlantic cod (Gadus morhua), and the eggs were fertilized in vitro at the Institute of Marine Research (IMR), Austevoll Research station. Fertilized eggs were transferred to indoor egg incubators, and maintained at 7 ± 1°C until transported to IMR's facilities in Bergen to begin the exposure studies at 5 days post fertilization (dpf), when the cardiac cone of cod embryos is visible. About 200 eggs were placed randomly in the prepared exposure beakers of either B[a]P or seawater (control group). There were four replicate beakers for each treatment and an additional beaker without eggs (water control) for each group. Exposure to B[a]P started at 5 dpf and lasted until 3 days post hatching (dph). The experiment took place in a cold room at 8°C, and the temperature in the water and in the room was monitored during the experiment. Dead eggs were removed during daily inspections from all the beakers to minimize bacterial contamination. Samples of several embryos and larvae were collected at 8 dpf, 10 dpf and 3 dph for gene expression and in situ hybridization analyses.

Water chemistry and body burden analyses
Water samples (1.0 mL) were taken at 5 dpf, 8 dpf, 11 dpf and 3 dph and liquid-liquid extracted with two 1-mL volumes of dichloromethane (SupraSolv®, Merck, Darmstadt, Germany). Extracts were dried with Na 2 SO 4 , and exchanged into 200 μL isopropanol (EMSURE®, Merck, Darmstadt, Germany). Twenty eggs were collected from each replicate jar at 8 dpf. These body burden samples were homogenized and extracted twice as in (Sørensen et al., 2019), with 50 % dichloromethane in n-hexane (SupraSolv®, Merck, Darmstadt, Germany) and subsequent clean-up with solid-phase extraction (Chromabond® silica, 3 mL, glass, Macherey-Nagel, Dueren, Germany). Water and body burden analyses for B[a]P and internal standard were performed on an Agilent 6890 gas chromatograph coupled to a 7010 triple quadrupole mass spectrometer as described in . Statistical differences in body burden concentration between control and B[a]P-exposed groups were analyzed using a Welch's t-test in R v1.2.1335 software (Crawley, 2012).

Quantitative polymerase chain reaction (qPCR) analyses
Total RNA was isolated from ten pooled samples of embryos or larvae from each beaker (n = 4) at 8 dpf, 10 dpf and 3 dph using the TRI Reagent® protocol (Sigma-Aldrich, St. Louis, Missouri, United States) as outlined by Chomczynski, 1993. 500 ng of total RNA was reverse transcribed to cDNA using the iScriptTM cDNA Synthesis Kit (Bio-Rad, California, USA) following the provider's protocol. Quantitative real-time polymerase chain reaction analyses (qPCR) were performed using SYBR Green Master I (Roche Diagnostics, Mannheim, Germany) and a CFX96 Touch Real-Time PCR Detection System (Bio-Rad, California, USA) as described in . Elongation factor 1 alpha (ef1a) and ubiquitin (ubi) were used as reference genes (Table S1) for normalization of gene expression across samples, and the geNorm software (Vandesompele et al., 2002) was used to calculate the normalization factor based on the geometric mean. The geNorm stability index M for the reference genes was 0.97. Relative induction of ahr1a, ahr2a, cyp1a and ahrrb genes (Table S1) between control and B[a]P-exposed groups was analyzed at the different sampling points using a Welch t-test on log transformed data in R v1.2.1335 software.

Tissue preparation for in situ hybridization
Cod embryos at 8 and 10 dpf and larvae at 3 dph were fixed in 4% paraformaldehyde buffered in 1x phosphate-buffered saline (PBS) solution (pH 7.4) for 48 h at 4°C for whole mount in situ hybridization. Whole mount embryos and larvae were briefly washed in 1x PBS, dehydrated in methanol, and stored in 100 % methanol at −20°C until use. Larvae for tissue sections in situ hybridization were fixed in 4% paraformaldehyde-buffered in 1x PBS for 48 h at 4°C, incubated in a solution of 25 % sucrose, 25 % Tissue Tek (Sakura Finetek, California, United States) and 50 % 1x PBS overnight at 4°C before they were oriented and embedded in Tissue Tek and kept at −80°C.

Molecular cloning
Cloning of ahr1a and ahr2a genes has been previously described in (Aranguren-Abadía et al., 2020). A partial sequence of cyp1a was amplified as a single fragment from cDNA prepared from juvenile cod heart using Taq DNA polymerase (Invitrogen, California, USA) and specific primers (Table S1). The cyp1a fragment was inserted into a pCR™4Blunt-TOPO® vector (Thermo Fisher, Massachusetts, USA), and DNA sequencing was performed at the University of Bergen Sequencing Facility.

Synthesis of ahr1a, ahr2a and cyp1a RNA probes
Preparation of digoxigenin (DIG)-labelled riboprobes for ahr1a, ahr2a and cyp1a were done following the manufacturer's instructions (Roche Diagnostics, Mannheim, Germany). In the synthesis of the riboprobes, PCR products was used as template for the reactions as described in (Thisse and Thisse, 2008) (Table 1) and the synthesized probes were precipitated by LiCl and EtOH together with tRNA (Roche Diagnostics, Mannheim, Germany). Further, ahr1a and ahr2a probes were designed in the less conserved coding sequence region in order to avoid cross-hybridization of probes between the paralogous genes.

Whole mount in situ hybridization
Whole mount in situ hybridization of embryos and larvae (n = 3 per gene and sampling point) were performed as described in (Valen et al., 2014). Briefly, embryos and larvae were rehydrated in methanol (75−25%) followed by dechorination of embryos. Pigmentation of embryos and larvae was removed by bleaching in 3% Hydrogen peroxide (H 2 O 2 )/0.5 % Potassium hydroxide (KOH) (Sigma-Aldrich, St. Louis, Missouri, United States). Prior to hybridization, the embryos and larvae were treated with Proteinase K (Promega, Wisconsin, United States). After hybridization, the tissue was thoroughly washed and treated with RNase A (Sigma-Aldrich, St. Louis, Missouri, United States) to remove unhybridized probe. Before applying the antibody (anti-DIG conjugated with alkaline phosphatase, Fab fragments (1:2000) (Cat. No 11093274910, Roche Diagnostics, Mannheim, Germany, RRID: AB_514497)) the embryos and larvae were incubated in 2% blocking solution (Roche Diagnostics, Mannheim, Germany) in 2x SSC with 0.05 % Triton X-100 (Sigma-Aldrich, St. Louis, Missouri, United States). The DIG-labelled probes were visualized by 4-nitro blue tetrazolium chloride (NBT) and 5-bromo-4-chloro-3-indolyl-phosphate system (BCIP) (Roche Diagnostics, Mannheim, Germany) and sense probes were used as a control for nonspecific DIG-probe labelling (data not shown).

Tissue sections in situ hybridization
In situ hybridization for arh1a, arh2a and cyp1a transcripts (sense and antisense) was performed on serial sections for n = 6 larvae per gene as described in (Ebbesson et al., 2011). By using a Cryostat (Leica CM 3050, Leica Microsystems, Wetzlar, Germany), 12 μm transversal sections were performed for each larvae and deposited into Superfrost Ultra Plus slides (Menzel-Gläser, Braunschweig, Germany). Proteinase K (Sigma-Aldrich, St. Louis, Missouri, United States) treatment was applied for 3.5 min and tissue sections were left 1 h at room temperature for drying prior to hybridization. Sections were treated with RNase (Roche Diagnostics, Mannheim, Germany) prior to incubation in antibody solution (1x maleate buffer + 1% block solution + 0,3% Triton X-100) with anti-digoxigenin alkaline phosphatase FAB-fragment (1:2000, (Roche Diagnostics, Mannheim, Germany)). Chromogen substrate (NBT/BCIP) (Roche Diagnostics, Mannheim, Germany) was applied for visualization of DIG-labelled probes and then mounted in 70 % glycerol for microscopy.

Imaging, craniofacial and cardiac analyses
Leica DMLB and M420 microscopes (Leica Microsystems, Wetzlar, Germany) with the camera Infinity 3 (Teledyne Lumenera, Ontario, Canada) and Image-Pro Premier 9.3 (Media Cybernetics, Rockville, Maryland, United States) were used for imaging of in situ hybridization samples. Adobe Photoshop CS5 (San Jose, California, United States) was used to adjust exposure, brightness and contrast in the pictures.
Larvae pictures and videos during the B[a]P-exposure were taken with an Olympus SZX-10 Stereo microscope equipped with a 1.2 Mp resolution video camera (Unibrain Fire-I 785c) controlled by BTV Pro 5.4.1 software (www.bensoftware.com). Larvae were immobilized in 3% methylcellulose and kept at 8°C using a temperature controlled stage (Brook Industries, Illinois, United States). Length of mandible, ethmoid plate, area of edema and the ventricular and atrial diastolic (D) and systolic diameter (S) used to estimate the fractional shortening (FS=(D-S)/D) were measured using Image J (Image J 1.52 t, National Institute of Health, Bethesda, Maryland, USA) with the object J plugin as described in (Sørhus et al., 2016). Differences in the length of jaw and ethmoid plate, edema area and FS between groups were assessed using one-way ANOVA and Tukey's multiple comparisons tests in R v1.2.1335 software.

Water chemistry and body burden analyses
Atlantic cod eggs were kept in normal seawater or seawater with B [a]P-infused silicone rods from 5 dpf until 3 dph. The concentration of B[a]P in water was measured at four time points, and body burden of B [a]P was measured after 3 days of exposure (8 dpf). Measured B[a]P concentration in water across all analyzed samples ranged between 0.3 and 37 μg/L with an average of 7.6 μg/L in seawater at 7°C (Fig. S1A).
The average body burden level in the B[a]P-exposed embryos was 14 ± 5 pg/embryo (range: 8-20); by wet weight, average body burden was 6.5 μg/kg wet weight (range: 3.7-9.2, Fig. S1B). Body burden in the B[a]P-exposed group was significantly greater (p = 0.01) than background levels present in the control group (0.36 μg/kg, Figure  S1B). Although body burden samples were only collected at 8 dpf, B[a] P is presumed to have been absorbed throughout the duration of the experiment.

Expression of ahr1a, ahr2a, cyp1a and ahrrb in eggs and larvae
Samples of embryos and larvae were collected at 8 dpf, 10 dpf and 3 dph, and the expression of ahr1a, ahr2a, cyp1a and ahrrb genes was assessed with qPCR analyses (Fig. 1). The expression of ahr1a did not change notably between the control and B[a]P-exposed groups at the different sampling points, but expression of ahr2a increased slightly at 10 dpf (p = 0.05) (Fig. 1A, B). cyp1a-expression was induced in B[a]Pexposed samples at 8 dpf (p = 0.01), 10 dpf (p = 0.001) and 3 dph (p = 0.00006), reaching close to a 40-fold average induction in transcript levels at 10 dpf and 3 dph compared to the normal seawater control samples (Fig. 1C). Expression of ahrrb did not change between the control and B[a]P-exposed groups at the different sampling points (Fig. 1D).

Whole mount in situ hybridization of Atlantic cod embryo and larvae
The specific localization of ahr1a, ahr2a, and cyp1a expression was assessed in cod embryo and larvae using in situ hybridisation. At 8 dpf, the expression of ahr1a was evident in the eye of both controls and B[a] P-exposed embryos ( Fig. 2A, B) (Fig. 2C, D). In contrast, differences in the ahr2a expression pattern was already detectable at this developmental stage (Fig. 2E, F), and ahr2a was only visible in the cardiovascular system of embryos exposed to B[a]P (Fig. 2G, H). An early effect of B[a]P-exposure on cyp1a expression was also visible, where its presence was detectable in the skin and vasculature of treated embryos (Fig. 2I, J). Expression of cyp1a was also localized in the heart and in the area where the liver develops in both control (Fig. 2L, M) and B[a] P-exposed embryos (Fig. 2O, P). There was not any visible expression of cyp1a in the eye of embryos from neither groups (Fig. 2K, N), as the staining present on the eye lens in B[a]P-exposed embryos corresponds to skin tissue (Fig. 2N).
At 10 dpf, ahr1a transcripts were present in the eye ganglion cell layer of embryos independent of their exposure regime (Fig. 3A, B, C, D). Notably, at this stage the expression of ahr2a was strongly visible in the cardiovascular system of exposed embryos (Fig. 3F, H) and still not detectable in embryos kept in normal seawater (Fig. 3E, G). Expression of cyp1a was also stronger in the skin and vasculature of B[a]P-exposed embryos (Fig. 3J) compared to the control group (Fig. 3I). cyp1a transcripts were also present in the heart and in the liver of embryos from both groups (Fig. 3L, M, O, P). No visible cyp1a expression was detected in the retinal layers of the control group (Fig. 3I, K) or the exposed group (Fig. 3J, N). However, some faint staining was seen in the lens in the control group, but this might be due to probe being trapped in the lens (Fig. 3K).
The specific expression of ahr1a in the eye was not clearly visible in the 3 dph whole mount larvae (Fig. 4A, B), but tissue sections revealed a persistent ahr1a expression in the ganglion cell and inner nuclear layers in larvae from both groups (Fig. 4C, D). At this stage, the Fig. 1. Expression of ahr1a, ahr2a, cyp1a and ahrrb in B[a]P-exposed embryos and larvae. Atlantic cod eggs and larvae were exposed to B [a]P in a passive-dosing system using infused silicone rods. Expression of ahr1a (a) ahr2a (b) cyp1a (c) and ahrrb (d) was measured with qPCR at 8 days post fertilization (dpf), 10 dpf and 3 days post hatching (dph). Log fold change in gene expression was calculated in comparison with the control group (normal seawater). Pool of 10 embryos or larvae (n = 4) were collected from each beaker, but some samples were compromised during laboratory analysis. The data are presented as normalized values (points) and mean (cross). Statistical differences in log transformed transcript levels between control and B[a]P-exposed samples at the different sampling points were assessed by using a Welch t-test in R v1.2.1335 software. Level of statistical significance is indicated with *(p ≤ 0.05), **(p ≤ 0.01) or *** (p ≤ 0.001).
expression of ahr2a was also intense in the cardiovascular system of B [a]P-exposed larvae (Fig. 4F) compared to the controls (Fig. 4E). In addition, expression of ahr2a was also detectable in the jaws, heart and fin nodes in larvae from both groups (Fig. 4G, H, J, K, M) and mainly in the liver area bordering towards the intestine of B[a]P-exposed individuals (Fig. 4L). In contrast to control larvae, there was a strong expression of cyp1a in the skin and vasculature of B[a]P-treated larvae, which made it difficult to capture the inner structures with conventional microscopy (Fig. 4N, O). However, in situ hybridization on tissue sections confirmed the presence of cyp1a in the heart and presumably the head kidney of both groups (Fig. 4Q, R, T, U) and in the liver and intestine of B[a]P-exposed samples (Fig. 4S). There was no expression of ahr2a or cyp1a in the liver from control larvae (Fig. 4I, P), but both ahr2a and cyp1a transcripts was observed in the liver of B[a]P-exposed larvae (Fig. 4L, S). Expression of those genes was also present in the heart of larvae from both groups, independent of the exposure regime (Fig. 4H, K).

Analyses of craniofacial and cardiac anomalies
No clearly visible malformations were observed in B[a]P-exposed 3 dph larvae, but differences in the length of jaw and ethmoid plate, edema area and FS between groups were measured for assessing putative deformities that can be produced after exposure to PAHs in early life stages (Table S2). Jaw lengths did not differ significantly between groups, and no consistent deviations in the length of the ethmoid plate Fig. 2. Whole mount in situ hybridization of control and B[a]P-exposed Atlantic cod embryos at 8 dpf. Expression of ahr1a, ahr2a and cyp1a was assessed in both B[a]P-exposed and control samples. All pictures show dorsal view of either the whole embryo (A, B, E, F, I and J) or detailed areas where gene expression was visible. Pictures of the eye (C, D, K, N), cardiovascular system (G, H), digestive system (L, O), and heart (M, P) areas were imaged. Gene expression was visualized by specific DIG-labelled RNA probes. e; eye, cs; cardiovascular system, h; heart, ds; digestive system, s; skin. Scale bars: 100 μm in the whole embryo pictures and 500 nm in the detailed areas pictures.
was revealed across the whole sample set (Table S2). Edema was not detected in 3 dph B[a]P-exposed larvae, nor any alterations in FS between control and B[a]P-exposed groups.

Discussion
In situ hybridization was used to study a possible subfunction partitioning of Ahr1a and Ahr2a in Atlantic cod by mapping the expression and localization of ahr1a, ahr2a and cyp1a during early development. Quantitative measurements showed an induction of the Ahr-target gene cyp1a at the different developmental stages assessed, confirming the activation of the Ahr-signaling pathway after B[a]P-exposure. An effect of B[a]P-exposure was also clearly visible in the skin of embryos and larvae where cyp1a was strongly expressed. Furthermore, the induction of ahr2a transcripts in the cardiovascular system of B[a]P-exposed embryos further supports an early effect of B[a]P on the Ahr-signaling pathway.
The purpose of the passive-dosing technique was to establish and maintain maximum aqueous B[a]P exposure. The measured concentrations of B[a]P in the exposure water had high variation but were reasonable in the context of solubility limits and anticipated deviations arising from experimental conditions. The published solubility of B[a]P in pure water is 1.52 μg/L at 25°C (Schwarzenbach et al., 2003). Our measurements were somewhat higher than the published value, despite Fig. 3. Whole mount in situ hybridization of control and B[a]P-exposed Atlantic cod embryos at 10 dpf. Expression of ahr1a, ahr2a and cyp1a was assessed in both B[a]P-exposed and control samples. All pictures show dorsal view of either the whole embryo (A, B, E, F, I and J) or detailed areas where gene expression was visible. Pictures of the eye (C, D, K, N), cardiovascular system (G, H), digestive system (L, O), and heart region (M, P) were imaged. Gene expression was visualized by specific DIG-labelled RNA probes. e; eye, cs; cardiovascular system, h; heart, ds; digestive system, s; skin. Scale bars: 100 μm in the whole embryo pictures and 500 nm in the detailed areas pictures. colder temperature and higher salinity, both of which should lower solubility. Binding of B[a]P to dissolved and particulate organic matter was the likely cause for the elevated and varying concentration measurements. The body burden analyses confirmed uptake of B[a]P into exposed fish embryos, while the actual body burdens were rather low. The uptake of B[a]P from aqueous media is limited by its very low solubility that is the combined result of a high hydrophobicity and high melting enthalpy (Kwon et al., 2016;Mayer and Reichenberg, 2006). However, the limited uptake is in line with previous passive-dosing exposure studies with Daphnia magna (Smith et al., 2010) and Lumbriculus variegatus (Agbo et al., 2013). Low body burdens can also be the results of biotransformation of B[a]P. At this developmental stage the liver bud is visible in Atlantic cod (Hall et al., 2004), and hepatic Cyp1a-mediated metabolism of B[a]P in fish has been demonstrated previously, including at early life stages during ontogenesis of the liver (Hornung et al., 2007;Kocan and Landolt, 1984;Stegeman et al., 1984;Stein et al., 1984;Varanasi et al., 1985). Low body burdens and efficient metabolism of B[a]P may also explain why the characeristic craniofacial and cardiac malformations often observed in early life stages of fish exposed to PAHs (Incardona, 2017) were not clearly visible in B [a]P-exposed cod larvae 3 dph.
Molecular mechanisms and physiological impacts of PAH-toxicity have been shown to differ among individual PAHs, and three possible modes of action have been described previously: Ahr-independent, Ahr- Fig. 4. In situ hybridization of whole mount and tissue sections prepared from control and B[a]P-exposed Atlantic cod larvae at 3 dph. Expression of ahr1a, ahr2a and cyp1a was assessed in both B[a]P-exposed and control samples. All pictures show dorsal view of either the whole larvae (A, B, E, F, N, O) or sectioned larvae. Pictures of the sectioned eye (C, D), jaws (G, K), heart (H, K,Q,T), liver (I, L, P, S), intestine (P, S), fin nodes (J, M), and presumably head kidney region (R, U) were imaged. Gene expression was visualized by specific DIG-labelled RNA probes. gcl; ganglion cell layer, inl; innear nuclear layer, onl; outer nuclear layer, j; jaws, h; heart, l; liver, fn; fin nodes, i; intestine, hk; head kidney. Scale bars: 100 μm in whole mount larvae and 500 nm in tissue sections except, 100 μm in G and K. dependent, and Cyp1a metabolism-dependent (Billiard et al., 2006;Clark et al., 2010;Incardona et al., 2005Incardona et al., , 2006. B[a]P is a 5-ring PAH and well-known environmental pollutant with thoroughly characterized cytotoxic, genotoxic, mutagenic and carcinogenic properties (Juhasz and Naidu, 2000;Uno et al., 2001;Verma et al., 2012). B[a]P toxicity has previously been shown to affect the developing heart and cardiovascular system of zebrafish embryos (Huang et al., 2012;Incardona et al., 2011) by modulating the activity of Ahr2 and induce cyp1a expression in the heart in an Ahr-dependent mode of action (Incardona et al., 2011). Similarly, expression of ahr2a was mainly visible in the cardiovascular system of exposed cod embryos and larvae, and in the heart of 3 dph larvae, indicating cardiotoxicity responses to B [a]P-exposure at the different developmental stages assessed. Notably, the heart is the first organ to become functional in fish, and this process is finalized in Atlantic cod around the golden eye stage. Around 8 dpf (40-somite stage) the tubular heart is present, but looping, demarcation of the atrium and ventricle chambers, and formation of the two minor chambers (sinous venosus and the bulbous arteriosus) occurs later in development (Hall et al., 2004). The heart is fully formed at 10 dpf (golden stage) as a result of an increase in the myocardial mass, which is produced by cell differentiation or proliferation (de Pater et al., 2009). Even though expression of ahr2a in the heart of B[a]P-exposed embryos was not clearly visible when using whole mount in situ hybridization, the presence of ahr2a in the cardiovascular system indicates an expression of this gene also in the heart of cod embryos. Moreover, the slight quantitatively increase in ahr2a mRNA expression observed in 10 dpf B[a]P-exposed embryos may be linked to the increase in myocardial cells during embryonic development.
Strong induction of cyp1a transcripts was revealed with quantitative measurements, and abundant expression of this gene was localized to the skin and vasculature in B[a]P-exposed individuals. Induction of cyp1a after B[a]P exposure has been previously observed in Atlantic cod and several other teleost species Carlson et al., 2004;Incardona et al., 2011;Van Veld et al., 1997;Yadetie et al., 2018). Induced Cyp1a protein expression in the skin, vasculature and cardiovascular system was also reported in zebrafish embryos exposed to high doses of B[a]P (Incardona et al., 2011) and other PAHs (Shankar et al., 2019). Furthermore, other exposure studies using different Ahr agonists, such as 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) or β-naphthoflavone (BNF), have also reported increased cyp1a gene expression and Cyp1a protein synthesis in the skin, vasculature, and cardiovascular system of zebrafish embryos and larvae (Andreasen et al., 2002;Goodale et al., 2012;Sugden et al., 2017). The skin represents the first line of defence in a water exposure system, so expression of cyp1a in the epidermis of Atlantic cod embryos and larvae is not surprising.
The expression of both ahr2a and cyp1a was induced in the liver of B [a]P-exposed 3 dph cod larvae. Similar observations were made by Andreasen et al., 2002, which observed expression of these two genes in the cardiovascular system and in organs such as skin, heart, liver and head kidney of zebrafish larvae after TCDD exposure (Andreasen et al., 2002). Taken together, our results indicate a role of Ahr2a in mediating toxicity to B[a]P in early life stages of Atlantic cod, which apparently involves Ahr2a-mediated induction of cyp1a gene expression. Developmental toxicology studies using zebrafish embryos have also previously demonstrated that Ahr2 is the main subtype involved in mediating toxic responses by inducing cyp1a expression (Goodale et al., 2012;Incardona et al., 2011;Prasch et al., 2003;Tanguay et al., 1999). We have recently characterized molecular and functional properties of the Atlantic cod Ahr1a and Ahr2a receptors, and our results showed transcriptional activation of both Ahr1a and Ahr2a in vitro after exposure to different Ahr ligands, including B[a]P. However, the high abundancy of ahr2a transcripts in the liver suggested that Ahr2a is the main protein involved in mediating xenobiotic responses in juvenile Atlantic cod (Aranguren-Abadía et al., 2020). Notably, the presence of ahr2a and cyp1a transcripts in the heart, liver, and presumably the head kidney, at the different developmental stages independent of the exposure regimes, indicates also a physiological role of these genes during early development of Atlantic cod. It has been hypothesized that Ahr's original function was to regulate the development of neural systems and sensory structures (Emmons et al., 1999;Hahn et al., 2017). In vertebrates, it has been shown that Ahr regulates different signalling pathways and physiological systems, such as the cardiovascular, reproductive and immune system (Esser and Rannug, 2015;Larigot et al., 2018;Mulero-Navarro and Fernandez-Salguero, 2016;Nebert, 2017). The expression of ahr2a in the jaws and fin nodes in both 3 dph B[a]Pexposed and control larvae suggests a physiological role of ahr2a in ontogenesis of these organs. Several studies have also demonstrated that zebrafish Ahr2 has a role in the development of the neuronal and reproductive systems (Garcia et al., 2018;Goodale et al., 2012;Souder and Gorelick, 2019). It was recently demonstrated that Ahr2-null zebrafish adults had malformed cranial skeletal bones and severely damaged fins (Garcia et al., 2018). Another study also showed craniofacial malformations and irregularities in the development of fins in Ahr2 knock-out zebrafish embryos (Souder and Gorelick, 2019). Hence, a role of Ahr2a in mediating the development of Atlantic cod's jaws and fins is not unlikely. In addition, a role of CYP1A in metabolizing endogenous compounds has also recently been postulated (Lu et al., 2020) and may explain the expression of cyp1a in control embryos and larvae tissues.
Importantly, the eye was the only tissue were ahr1a transcripts were detected at all developmental stages assessed. Staining was strong in the ganglion cell layer already in 8 dpf embryos, and expression of ahr1a was also present in the eye inner cell layer of 3 dph larvae. In juvenile Atlantic cod, ahr1a expression was found in liver, brain and gill tissue, in addition to the eye (Aranguren-Abadía et al., 2020). A physiological role of the zebrafish ahr1b in embryonic development was postulated after high levels of ahr1b transcripts were detected in zebrafish embryos (Karchner et al., 2005). Recently, Karchner et al. have identified zfAhr1b protein in the inner plexiform and outer plexiform layers in the developing eye of zebrafish, and a down-regulation in the expression of genes involved in phototransduction in zfAhr1b knockdown embryos . Different localization between cod ahr1a transcripts and zebrafish Ahr1b proteins in the eye layers may be explained by a translocation of Ahr1b to the plexiform layers subsequently to protein synthesis. Importantly, our results indicate that Atlantic cod's Ahr1a protein may be involved in embryonic development of the eye.

Conclusions
The spatio-temporal expression patterns of ahr1a, ahr2a, and cyp1a were mapped during early development of Atlantic cod. Different expression levels and localization of ahr2a and cyp1a transcripts were observed between control and B[a]P-exposed embryos and larvae at all developmental stages assessed. In B[a]P-exposed embryos and larvae the expression of ahr2a was mainly localized in the cardiovascular system, whereas cyp1a was mainly induced in the skin and vasculature. The presence of ahr2a and cyp1a in the liver of B[a]P-exposed larvae further supports a role of Ahr2a in mediating responses to xenobiotics in early life stages. Furthermore, the expression of ahr1a in the eye of both embryos and larvae, and the presence of ahr2a in the jaws and find nodes of larvae, also suggest roles of these proteins in ontogenesis during early development. Importantly, our data support that a subfunction partitioning of Ahr1a and Ahr2a has occurred in Atlantic cod.

Author contributions
LA-A, CED, ES, SM were involved in experimental design and sampling. LA-A performed RNA extraction and qPCR analyses, probedesign and synthesis, whole mount in situ hybridization analyses, whole mount and tissue section samples imaging, craniofacial and cardiac measurements, and statistical analyses. CED analyzed water chemistry and body burden samples. ME contributed with whole mount in situ hybridization, and NG and VT performed in situ hybridization on tissue sections. ES contributed with the craniofacial and cardiac analyses, PM with expertise in passive-dosing techniques and TON and VT with cyp1a plasmid preparation. LA-A, CED, and OAK wrote the manuscript, with LA-A as the lead author. All authors contributed with interpretation of the results and revision of the manuscript.

Funding
This work was supported by Research Council of Norway (iCod 2.0: Integrative environmental genomics of Atlantic cod project (project no. 244564)), (EGGTOX: Unraveling the mechanistic effects of crude oil toxicity during early life stages of cold-water marine teleosts (project no. 267820)), (The effect of narrow banded LED light on development and growth performance (project no. 254894)), the Center for Digital Life Norway (dCod 1.0: decoding systems toxicology of Atlantic cod project, (project no. 248840) and by the Norwegian Research Center (NORCE).

Declaration of Competing Interest
The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.