|Home | About | Journals | Submit | Contact Us | Français|
Knowledge of the complement of cytochrome P450 (CYP) genes is essential to understanding detoxification and bioactivation mechanisms for organic contaminants. We cloned three new CYP1 genes, CYP1B1, CYP1C2 and CYP1D1, from the killifish Fundulus heteroclitus, an important model in environmental toxicology. Expression of the new CYP1s along with previously known CYP1A and CYP1C1 was measured by qPCR in eight different organs. Organ distribution was similar for the two CYP1Cs, but otherwise patterns and extent of expression differed among the genes. The AHR agonist 3,3′,4,4′,5-pentachlorobiphenyl (PCB126) (31 nmol/g fish) induced expression of CYP1A and CYP1B1 in all organs examined, while CYP1C1 was induced in all organs except testis. The largest changes in response to PCB126 were induction of CYP1A in testis (~700-fold) and induction of CYP1C1 in liver (~500-fold). CYP1B1 in liver and gut, CYP1A in brain and CYP1C1 in gill also were induced strongly by PCB126 (>100-fold). CYP1C1 expression levels were higher than CYP1C2 in almost all tissues and CYP1C2 was much less responsive to PCB126. In contrast to the other genes, CYP1D1 was not induced by PCB126 in any of the organs. The organ-specific response of CYP1s to PCB126 implies differential involvement in effects of halogenated aromatic hydrocarbons in different organs. The suite of inducible CYP1s could enhance the use of F. heteroclitus in assessing aquatic contamination by AHR agonists. Determining basal and induced levels of protein and the substrate specificity for all five CYP1s will be necessary to better understand their roles in chemical effects and physiology.
Cytochrome P450 (CYP) enzymes catalyze oxidative metabolism of thousands of drugs, environmental pollutants, and endogenous compounds. Environmental pollutants such as halogenated hydrocarbons, polycyclic aromatic hydrocarbons (PAH), herbicides, and pesticides (Nebert and Russell, 2002) include many substrates of mammalian CYP1s (CYP1A1, CYP1A2, CYP1B1), and of fish CYP1As (e.g., Schober et al., 2006). While metabolism often results in detoxification, the action of CYP1 enzymes also can generate toxic metabolites that contribute to increased risks of cancer, birth defects, and other toxic effects (Nebert and Karp, 2008). Expression of mammalian CYP1A1 and fish CYP1As can be induced strongly by PAH, planar polychlorinated biphenyl (PCB), dibenzo-p-dioxin (PCDD), and dibenzofuran (PCDF) congeners, and some natural products, via activation of aryl hydrocarbon receptor (AHR) (Hahn, 2002). These features have led to widespread use of CYP1A gene expression as a marker of environmental exposure to AHR agonists in humans and wildlife (e.g., Stegeman, 1986; Fujita et al., 2001; Lambert et al., 2006).
Most fish have only one CYP1A gene and one CYP1B1 gene (Goldstone et al., 2007). Fish also have two CYP1Cs, the paralogous CYP1C1 and CYP1C2 (Godard et al., 2005). Like CYP1A and CYP1B1, the CYP1Cs are induced to varying degrees by AHR agonists in zebrafish (Jönsson et al., 2007a; Jönsson et al., 2007b). More recently we have identified a fifth CYP1 gene in teleosts, CYP1D1 (Goldstone et al., 2007; Goldstone and Stegeman, 2008). Zebrafish CYP1D1 and CYP1A share a relatively high percent identity and have similar gene structures. However, neither PCB126 nor TCDD induced transcription of zebrafish CYP1D1. CYP1D1 protein is expressed in zebrafish liver, and heterologously expressed CYP1D1 is catalytically active with ethoxyresorufin, albeit at significantly lower rates than CYP1A (Goldstone et al., 2009). It is possible that chemicals and receptors other than AHR are involved in regulating these various CYP1s, and that they act on distinct sets of substrates, which could vary among species.
To date, all five CYP1s have been examined in detail only in zebrafish. We sought to determine if the identity of the full suite of CYP1 genes occurs and is similarly regulated in the Atlantic killifish Fundulus heteroclitus. This species is one of the most abundant estuarine fishes along the Atlantic coast of North America, and frequently is used in toxicological studies, including developmental toxicology and environmental carcinogenesis (Burnett et al., 2007). Biological and ecological features (small size, rapid development, small home range, tolerance of varied conditions) and adaptation to high levels of contaminants have spurred interest in this species as an alternative non-mammalian vertebrate model species (Burnett et al., 2007; Matson et al., 2008).
Regulation of CYP1A-like activities has been studied in F. heteroclitus for 30 years (e.g., Stegeman, 1978) and CYP1A was cloned a decade ago (Morrison et al., 1998). More recently, a killifish CYP1C1 was reported (Wang et al., 2006). Here, we report on the identification and cloning of the full-length sequences of three new F. heteroclitus CYP1 genes, CYP1B1, CYP1C2, and CYP1D1. Basal expression and regulation by the AHR agonist PCB126 were compared in different organs for the three new genes, and for CYP1A and CYP1C1. It appears that orthologous CYP1 genes are similarly expressed in Fundulus and in zebrafish. Identifying substrates of the CYP1s in this model will help to achieve a general understanding of their involvement in chemical effects and their roles in endogenous functions, and may contribute to understanding the resistance to AHR agonist toxicity seen in some populations.
Killifish were obtained from Scorton Creek (Massachusetts, USA) in May 2008, using minnow traps. Scorton Creek has been used as a reference site for studies of F. heteroclitus for years (Bello et al., 2001). Fish were acclimated at 20 °C in flowing seawater for one month and were fed twice a day with Omega One Freshwater Flakes™ during the acclimation period. The procedures used in these experiments were approved by the Animal Care and Use Committee of the Woods Hole Oceanographic Institution.
Liver, brain and eye were dissected from one randomly selected untreated fish. Total RNA was isolated using Aurum™ Total RNA Fatty and Fibrous Tissue Kit (Bio-Rad Laboratories Inc., Hercules, CA), which includes elimination of genomic DNA by DNAse treatment. The RNA quantity and quality was determined spectrophotometrically (Nanodrop ND 1000; NanoDrop Technologies, Wilmington, DE). cDNA was synthesized from 2 μg of total RNA, using the Omniscript Reverse Transcriptase kit (Qiagen Inc., Valencia, CA), anchored oligo(dT) primer (MWG Biotech, Inc., High Point, NC) and RNasin RNase inhibitor (Promega Corp., Madison, WI).
Degenerate primers were designed using highly conserved regions of CYP1B1 and CYP1D1 sequences from other fish species (Supplemental Table 1), avoiding conserved regions present in other CYP1 subfamilies. Gene-specific primers for CYP1C2 were designed against a sequence fragment in GenBank. PCR reactions were carried out for CYP1B1 and CYP1C2 using brain cDNA and for CYP1D1 using liver cDNA. PCR products were resolved on a 1% agarose gel and then isolated, ligated into the pGEM-T Easy Vector (Promega), and transformed into E.coli (TOP 10 Kit, Invitrogen). Plasmids were purified from cultures of positive clones (QiaPrep™, Qiagen) and were sequenced (MWG Biotech).
The 5′ and 3′ ends of CYP1B1, CYP1C2 and CYP1D1 were obtained by rapid amplification of cDNA ends (RACE) with the BD Smart™ RACE cDNA Amplification Kit (Clontech) using the RACE kit and gene-specific primers described in Supplemental Table 1. Gel-purified 3′ and 5′ RACE products were cloned and sequenced as described above.
Nucleotide sequences were translated and aligned with other CYP1 family members as before (Jönsson et al., 2007b; Goldstone et al., 2009). Predicted medaka (Oryzias latipes) and stickleback (Gasterosteus aculeata) CYP1 gene sequences were derived from the Ensembl genomes (Release 49). Phylogenetic trees were constructed by analyzing predicted amino acid sequences using maximum likelihood (RAxML 7.0.3; (Stamatakis, 2006b)), and Bayesian methods (MrBayes 3.1.2 (Ronquist and Huelsenbeck, 2003)). Regions of alignment uncertainty were excluded from phylogenetic analysis (Kreil and Ouzounis, 2003) by automatic masking using a custom-written script. The WAG-CAT model of amino acid substitution with a gamma distribution of substitution rates (PROTMIXWAG) was used in all likelihood analyses, based on likelihood tests using RAxML (Stamatakis, 2006a). MrBayes estimates posterior probabilities using Metropolis-Hastings coupled Monte Carlo Markov chains (MC3). MC3 estimates were performed with the WAG model of amino acid substitution and prior uniform gamma distributions approximated with four categories (WAG+I+Γ). Four incrementally heated, randomly seeded Markov chains were run for 3×106 generations, and topologies were sampled every 100th generation. Burn-in value was set to 106 generations. Putative functional domains (i.e., representing substrate recognition sites) were evaluated for similarity to domain sequences in orthologous CYP1 genes in other species. The amino acid sequence similarities between various CYP1s were plotted using GCG (v. 10.3; Accelrys, San Diego, CA). Similarity scores were calculated using the BLOSUM62 amino acid similarity matrix (Henikoff and Henikoff, 1992).
Twenty male fish (3-5 g whole body weight) were acclimated for 24 hr in recirculating, filtered and aerated water in two 10-gallon aquaria (10 fish per aquaria) at 18 °C. After acclimation the fish were weighed and 10 fish were injected intraperitoneally with 3,3′,4,4′,5-pentachlorobiphenyl (PCB126) dissolved in DMSO at the dose of 31 pmol/g fish (10 μg/kg fish). Ten fish were also injected with an equivalent volume of DMSO alone. Previous studies in our laboratory using scup (Stenotomus chrysops), showed that injection of this PCB126 dose caused high-level induction in gene expression and catalytic function (EROD) of CYP1A with absence of mortality (unpublished). The same PCB126 dose was also injected in pink snapper (Pagrus auratus) in another study, which gave similar results in CYP1A induction (Tugiyono, 2002). At 48 hrs after injection, control and PCB126 treated fish were killed by cervical transection and liver (n = 10), gill (n = 8), gut, heart, kidney, eye, brain, and testis (n = 4) of individual fish were dissected and immediately placed in RNAlater (Ambion). The samples were held for 24h at 4 °C, and then stored at -20 °C. Previous studies in our lab showed higher induction of CYP1A gene expression in fish exposed to AHR agonists using 48hs exposure. This was also observed by Kim et al. (2008) who tested a variety of exposure times, and analyzed CYP1A in different pufferfish organs.
Total RNA was extracted and cDNA was synthesized as described above. Gene-specific primers for the new F. heteroclitus CYP1s and for F. heteroclitus CYP1A, CYP1C1, and β-actin designed with Primer3 (Rozen and Skaletsky, 2000) were obtained from MWG Biotech (Primer sequences are shown in Table 1). Real-time PCR was performed using iQ SYBR Green Supermix (according to the manufacturer’s instructions) and an iQ Real-Time PCR Detection System (Bio-Rad). For each sample, gene expression was analyzed in triplicate with the following protocol: 95°C for 3 min and 40 cycles of 95°C for 15 s and 62°C for 1 min. Melt curve analysis was performed on the PCR products at the end of each PCR run to ensure that a single product was amplified. The EΔct method was used to compare the expression levels of the different CYP1s within a given organ, and to calculate changes in fold-induction in response to PCB126 treatment. The efficiency of the PCR reactions for each gene was calculated using the standard curves generated from dilutions (10 2 – 10 9 molecules) of plasmid (pGEM-T Easy Vector) containing fragments of a given target gene (Schmittgen and Livak, 2008). Jönsson et al. (2007b) compared data normalized to different house-keeping genes (β-actin or ARNT) using the EΔct method and showed that results differ between these normalizations. Organ-specific differences in Ct values for β-actin and other reference genes in fish have been shown before (McCurley and Callard, 2008), and produce conflicting qPCR results when using the Δct housekeeping — ct target normalizing method. Therefore, to compare relative expression levels between different organs, molecule numbers for CYP1s and β-actin were calculated based on standard curves and relative CYP1 transcript abundance was determined for a specific amount of total RNA from each organ using β-actin only as an efficiency control to normalize individual samples within a specific organ, but not among different organs.
Equality of variance was tested, accepting P>0.05 (Bartlett’s test) (Prism 4, GraphPad Software Inc., San Diego, CA). Data were logarithmically transformed to improve equality of variances. The differences among transcript levels for CYP1s in different organs were evaluated using one-way analysis of variance (ANOVA, P<0.05), followed by Tukey’s test for multiple comparisons (Prism 4, GraphPad Software Inc., San Diego, CA). Differences between the means of control and PCB126 treated groups were analyzed using Student’s t-Test and data are presented as fold difference between treated and control groups.
Using degenerate primers based on CYP1B1 and CYP1D1 in other teleosts, we were able to amplify partial sequences (about 700 bp) from F. heteroclitus. Aligning the amino acid sequences predicted from these cloned PCR products with other fish CYP1 amino acid sequences resulted in preliminary classification of the new Fundulus sequences as CYP1B1 and CYP1D1.
The alignment of fish CYP1s included a predicted translation of a 588 bp sequence obtained from GenBank (Accession AF235140) that was annotated as F. heteroclitus CYP1B1. However, alignment of the predicted amino acid sequence with other CYP1 sequences indicated that this 588 bp sequence was likely a CYP1C2, and not a CYP1B1. This putative CYP1C2 sequence and the cloned fragments of the other two genes were used as starting points for full-length sequencing using PCR reactions primed with RACE and gene specific primers. Full-length sequences were obtained for all three genes (GeneBank accession FJ786959, FJ786960 and FJ786961, respectively).
The open reading frames for the three new Fundulus genes give translated proteins of 537, 524 and 535 amino acids, 10-15 amino acids longer than the corresponding zebrafish CYP1B1, CYP1C2, and CYP1D1, respectively (Supplemental Figure 1). The full-length killifish sequences display 61%, 69% and 67% pair-wise identity to CYP1B1, CYP1C2 and CYP1D1 of zebrafish, respectively, increasing to 69%, 74%, and 75% identity when ambiguously-aligned regions of a large set of CYP1s are masked (Table 2). Notably, our annotation of these subfamilies and genes is in agreement with the CYP classification based on the amino acid percent identities suggested by (Nelson et al., 1996), accepting identities higher than 55 % for the same subfamily.
Phylogenetic analysis of the deduced amino acid sequences for the five F. heteroclitus CYP1s was done together with members of the CYP1 family from zebrafish and other selected species. These include medaka and stickleback, where five CYP1 sequences were found in their genomes. Figure 1 shows that the killifish sequences clustered with the other teleost CYP1s, with the CYP1As and CYP1D1s appearing in one clade, and the CYP1B1s and the CYP1Cs occurring in another. In all cases, the F. heteroclitus CYP1s clustered most closely with the predicted CYP1 orthologues of medaka. Interestingly, as in previous analyses (Goldstone et al., 2007; Goldstone et al., 2009), in this analysis zebrafish CYP1C1 and CYP1C2 clustered together rather than with their respective orthologues. This general result may be due to gene conversion, as seen in avian CYP1As (Goldstone and Stegeman, 2006), although the taxonomic sampling of the CYP1Cs is not yet sufficient to conclude this with any certainty.
In general, the corresponding CYP1s in Fundulus and zebrafish exhibit high degrees of amino acid sequence similarity (Supplemental Figure 2). However, this similarity varied along the length of the proteins, with the lowest similarities (BLOSUM62-based score) observed in the membrane anchor, D helix, and the disordered H-I loop regions. The substrate recognition sites (SRS) of the orthologous zebrafish and Fundulus CYP1s are 74-84% identical, with SRS 2 and 3 displaying the largest number of differences (Figure 2). Comparisons of F. heteroclitus CYP1A and CYP1D1 sequences along the length of the proteins showed a marked likeness between CYP1A and CYP1D1 in SRSs 4 and 5 and a marked difference in SRSs 2 and 3 (Figure 3). These results are similar to the variation in sequence similarities observed between zebrafish CYP1A and CYP1D1 (Goldstone et al., 2009).
Figure 4 shows the differences in expression levels among the CYP1 genes in a given organ. Relative levels of expression are calculated and expressed as Δ ct β-actin - ct target. In liver, heart, kidney, gut, eye and gill, CYP1A was the highest expressed transcript. In liver, CYP1A was ~300 times higher than CYP1D1, ~1,300 times higher than CYP1B1 and CYP1C1, and ~30,000 times higher than CYP1C2. In brain, CYP1B1 and CYP1D1 were the most highly expressed transcripts. In testis, CYP1C1 was the most highly expressed transcript, ~1,000 and 3,000 times higher than CYP1A and CYP1B1, respectively. CYP1C2 was expressed at the lowest levels among the five CYP1 genes in most of the organs examined (liver, heart, kidney, eye, brain and kidney).
Figure 5 shows relative transcript levels of the five CYP1s measured by qPCR, comparing the levels of a given gene in the eight organs we sampled from control fish: liver, kidney, gut, heart, gill, eye, brain and testis. The analysis shows that the highest levels of CYP1A transcript were observed in liver. Levels of CYP1B1 transcript were similarly high in brain, heart, gill, eye and kidney. Earlier studies observed differences in the basal levels of two reference genes, β-actin and ARNT2, among the different organs (Jönsson et al., 2007b). Thus, in order to avoid erroneous comparisons of specific CYP1 transcripts between different organs based on a given reference gene, we calculated transcript abundances based on standard curves. Individual sample correction for β-actin levels were organ-specific to allow for CYP1 transcript comparison among organs. The highest levels of both CYP1C1 and CYP1C2 were in kidney, and of CYP1D1 were in brain and kidney. The lowest levels of CYP1A were observed in testis, brain and eye; CYP1B1 in testis, gut and liver; CYP1C1 and CYP1C2 in liver and gill; CYP1D1 in liver and gut. CYP1C1 and CYP1C2 showed generally similar patterns of basal expression levels among organs.
No mortality was observed in F. heteroclitus injected with PCB126 or the carrier DMSO. PCB126 induced the expression of CYP1A and CYP1B1 in all eight organs we examined (liver, heart, kidney, gut, eye, brain, gill and testis) (Figure 6). The most substantial changes in CYP1 expression in response to PCB126 were in liver, where CYP1C1 was induced ~500-fold and CYP1B1~200-fold, and in testis, where CYP1A was induced ~700-fold. CYP1B1 in liver and gut, CYP1A in brain and CYP1C1 in gill also were induced strongly (~100-fold in each case).
CYP1C1 also was induced in all organs except testis. Although CYP1C1 and CYP1C2 had similar transcript profiles in the different organs of control fish, CYP1C2 was much less responsive to PCB126 compared with CYP1C1. Strikingly, in contrast to the other four genes, CYP1D1 was not significantly induced by this dose of PCB126 in any of the organs examined.
With the cloning and sequencing of three new CYP1 genes the CYP1 family in F. heteroclitus is expanded to four subfamilies and five genes, CYP1A, CYP1B1, CYP1C1, CYP1C2 and CYP1D1. Previously we established that the same set of CYP1 subfamilies and genes occurs in zebrafish, and that the five zebrafish CYP1s are expressed at the level of transcription (Jönsson et al., 2007a; Jönsson et al., 2007b; Goldstone et al., 2009). The five genes in Fundulus appear to be orthologues of those in zebrafish, and likewise all five are expressed in many organs.
Molecular phylogenetic analysis shows that Fundulus CYP1A and CYP1D1 are grouped together in one clade, while CYP1B1 and the CYP1Cs are in another. This result is fully consistent with our prior observation that the CYP1Cs and CYP1Bs are sister subfamilies occurring in one monophyletic clade (Goldstone et al., 2007; Jönsson et al., 2007b) and that the CYP1As and CYP1Ds are in another (Goldstone et al., 2009). The common molecular phylogeny for the CYP1 genes in several species thus supports the hypothesis that the CYP1As and CYP1Ds diverged from a common CYP1A/CYP1D ancestor, and the CYP1Bs and CYP1Cs from a common CYP1B/CYP1C ancestor (Goldstone et al., 2007; Goldstone et al., 2009).
Killifish CYP1C1 and CYP1C2 are very closely related phylogenetically. A similarly close relationship occurs between CYP1C1 and CYP1C2 genes in fugu (Godard et al., 2005), and in zebrafish, where the two genes are located immediately adjacent to one another on chromosome 17 (Jönsson et al., 2007b). These observations suggest that the two CYP1C paralogs resulted from an independent duplication subsequent to a divergence of the CYP1B and CYP1C lines. Further, the observations are consistent with the idea that this gene duplication occurred subsequent to the branching of fishes from the vertebrate line, and that it occurred more recently than the whole genome duplication thought to have taken place in the fish lineage.
The phylogenetic analyses also show that the killifish CYP1 sequences are more closely related to the CYP1 sequences in medaka and stickleback than to those in zebrafish. These similarities could be expected, since zebrafish is in the superorder Ostariophysi while medaka, stickleback and Atlantic killifish are in the superorder Acanthopterygii. These two superorders were separated about 290 million years ago, while orders Beliniformes (medaka) and Cyprinodontiformes (Atlantic killifish) were separated more recently, around 153 million years ago (Steinke et al., 2006).
When compared to other CYP1s, higher levels of CYP1A transcript were observed in most of the organs analyzed, except brain and testis. Relatively higher levels of CYP1A were also observed in the organs of abdominal cavity (e.g. liver, kidney and gut) in zebrafish, and could be associated with the role of these organs in nutrient uptake and processing of body waste products, e.g., detoxification of endogenous metabolites and food-derived AHR agonists (Jonsson et al., 2007b). Although CYP1A was the most abundant transcript in eye and heart, CYP1B1 and CYP1C1 were also expressed in substantial levels in these organs. Yin et al. (2008) showed that the basal level of CYP1B1 in zebrafish embryos is regulated by an AHR2-independent pathway, and that CYP1B1 is not directly involved in pericardial edema generated by dioxin toxicity. It is well known that CYP1B1 is important for normal eye development in mammals (Choudhary et al., 2006) and that mutations in this gene are correlated with glaucoma in human populations (Ohtake et al., 2003). The physiological importance of CYP1B1 and CYP1C1 basal levels in the adult fish eye and heart remains to be elucidated.
Similar to the present study, previous studies have shown greater levels of CYP1B1 and CYP1D1 expression in brain of adult zebrafish compared to the other CYP1s (Jönsson et al., 2007b; Goldstone et al., 2009). In developing zebrafish CYP1B1 expression was seen in the brain before hatching (Yin et al., 2008). In rat brain as well the basal levels of expression of CYP1B1 were higher than CYP1As, in four distinct regions of the brain (Desaulniers et al., 2005). While both CYP1B1 and CYP1D1 are expressed in brain the significance of this expression to brain function is unknown. The relatively higher levels of CYP1B1 expression in brain suggest that these CYP1s could play particular roles in this organ in vertebrates including fish and mammals. The same may be true for CYP1D1 in species where this gene is expressed (see below).
High levels of all five CYP1s were observed in kidney. CYP1C1, CYP1C2 and CYP1D1, were particularly high when compared to their levels in other organs. There could be an important function for those CYP1s in detoxification; high levels of gene expression for many phase-I, II and III biotransformation enzymes, as well as receptors involved in the regulation of those genes, are observed in the kidney and participate in the elimination of many xenobiotics, drugs and endogenous compounds (Xu et al., 2005).
Comparing the levels of expression of each CYP1 in different organs of killifish showed similar patterns for the two CYP1Cs, with higher levels in kidney, brain and testis, and lower levels in liver and gill. The relative levels of expression of CYP1C1 in kidney, liver, testis and eye observed by Wang et al., (Wang et al., 2006), in a different population of F. heteroclitus, are similar to the CYP1C results presented here. However, while there are similarities in the organ-specific patterns of expression of the CYP1Cs, CYP1C2 was expressed at much lower levels than CYP1C1. Thus, the two CYP1Cs could be regulated by similar organ-specific pathways, but may yet function differently in a given organ.
The determination that these five CYP1s are expressed in Fundulus as they are in zebrafish is important, perhaps especially so in the case of CYP1D1. This is because in humans CYP1D1 is a pseudogene (CYP1D1P) (Goldstone et al., 2009) and it appears to be a pseudogene in some other species as well (unpublished data), and thus, finding a CYP1D1 gene does not necessarily mean that it will be expressed. The CYP1D1 locus may have been lost from the genome of still other species, including some fish; we have been unable to detect a CYP1D in pufferfish genomes. In some species CYP1D1 functions may be accomplished by other CYPs, possibly other CYP1s. It will be interesting to determine whether CYP1D1 has unique substrate specificities.
Generally, it appears that there are catalytic similarities among fish CYP1As. The substrate specificities of the other CYP1s in Fundulus are not yet known, and are only poorly known in zebrafish. We have expressed the five zebrafish CYP1s in yeast, and determined that the zebrafish CYP1Cs and CYP1D1 act on some of the same alkoxyresorufin substrates that are oxidized by CYP1A (Goldstone et al., 2009 and unpublished). In the eel Anguilla japonica, CYP1C1 appears to metabolize the CYP1A substrates 7-ethoxyresorufin and 7-ethoxycoumarin, although with lower enzymatic activity than CYP1A. Interestingly, there were differences between the eel CYP1A and CYP1C1 in the products formed from one substrate; eel CYP1C1 produced two products from flavanone whereas CYP1A produced just one (Uno et al., 2008).
Studies with heterologously-expressed enzymes can indicate substrate specificities of the different CYP1s. However, to understand the role of a particular CYP in a given organ will require determining the levels of protein expressed, and contributions to substrate turnover by use of specific inhibitors, or inhibitory antibodies. We have obtained antibodies to zebrafish CYP1D1 and have determined that CYP1D1 protein is expressed in zebrafish liver (Goldstone et al., 2009). Unfortunately, the antibodies to zebrafish CYP1D1 do not recognize a protein in Fundulus liver (unpublished data). Determining the catalytic function of the various CYP1s with multiple substrates will be essential to establishing the physiological roles of these enzymes. At present this would be merely a matter of speculation. However, the expression patterns suggest to us that these enzymes will be shown to have distinct physiological roles in fish.
The five CYP1s in killifish differed in their responses to treatment of the animals with the potent AHR agonist PCB126. The observation that CYP1A is induced in all organs was expected. Previous studies have shown that CYP1A in fish is induced to high levels of expression in detoxification organs (e.g., liver, gastrointestinal tract, gill and kidney), and that CYP1A is induced in some cell types, often in endothelium, in all organs (e.g., Smolowitz et al., 1992). Immunohistochemical studies have also shown that CYP1A protein levels are strongly induced in all organs of Fundulus exposed to AHR agonists (VanVeld et al., 1997). The present study also shows that CYP1B1 was induced in all organs examined. The most significant changes in CYP1 expression in response to PCB126 were induction of CYP1A (~700-fold) in testis and induction of CYP1C1 (~500-fold) in liver. CYP1B1 in liver and gut, CYP1A in brain and CYP1C1 in gill also were strongly induced (>100-fold). Low basal levels of the respective CYP1 gene expression, together with a strong induction by AHR agonists, could explain the higher fold-increases over control values observed in some organs.
In contrast to CYP1A, CYP1B1 and CYP1C1, expression of CYP1C2 transcript was weakly induced, and only in three of the organs analyzed (liver, heart and gill). The lesser responsiveness of CYP1C2 to a potent AHR agonist is reminiscent of the results obtained with zebrafish, in which the responsiveness of CYP1C2 to PCB126 was largely absent from adults (Jönsson et al., 2007a; Jönsson et al., 2007b). The similarity of results observed in killifish and zebrafish, even considering that there were different exposure routes for PCB126 in the experiments with adult fish, suggest a silencing of the induction response of CYP1C2 in adults of these fish species. Distinct from the other CYP1s, CYP1D1 was not induced in any of the eight organs we examined here. In like fashion, there was no induction of CYP1D1 in zebrafish adults treated with PCB126 or TCDD (Goldstone et al., 2009).
It might be argued that response of CYP1s in Fundulus embryos could differ from that in adults. In zebrafish, the only notable difference between adult and embryonic responses to PCB126 was that CYP1C2 was induced in zebrafish embryos. However, in zebrafish the responsiveness of CYP1C2 to PCB126 decreased with developmental age. The developmental pattern of expression and the response of the five CYP1 genes to AHR agonists during embryonic development in Fundulus is under investigation.
The mechanism involved in the lesser induction response of CYP1C2 in adult fish is not known. However, the lack of induction of CYP1D1 in zebrafish was suggested to be related to a lack of functional AHR response elements (AHREs), which are binding sites for the AHR/ARNT in the promoter regions (Goldstone et al., 2009). In zebrafish, there are only two putative AHREs (of unproven function) in the promoter of zebrafish CYP1D1, in contrast to the 22 putative AHREs and three proven functional AHREs for CYP1A (ZeRuth and Pollenz, 2007; Goldstone et al., 2009). A variety of other response elements were identified in the promoter region of zebrafish CYP1D1 (Goldstone et al., 2009) and thus the possibility of induction via other receptors and agonists still remains to be elucidated, both for the CYP1Cs and CYP1D1 in Fundulus and in zebrafish.
The induction of CYP enzymes in fish liver was first suggested as an indicator of aquatic contamination in the 1970s (e.g., Payne, 1976). Since then, many studies have shown that CYP1As in vertebrate liver (often measured by activity assay and protein detection by Western blot) are strongly induced by some organic contaminants that represent risk for human and wildlife (e.g., PAHs, coplanar PCBs, polychlorinated dibenzofurans, and dibenzodioxins) (Bucheli and Fent, 1995). The results of the present study, along with the results of Wang et al., (Wang et al., 2006), suggest that in addition to CYP1A, other CYP1s, notably CYP1B1 and CYP1C1, have potential to be additional sensitive biomarkers of exposure to AHR agonist contaminants in F. heteroclitus and likely other fishes. Lesser responsiveness by CYP1C2 in adults and the lack of CYP1D1 induction by potent AHR agonists in two different species suggests that these two genes would not be suitable markers for exposure to such chemicals. Nevertheless, CYP1C2 and CYP1D1 may still be involved in the toxicity of AHR agonists or other chemicals, if such chemicals are substrates for these enzymes, or if these enzymes have important endogenous substrates and are inhibited by xenobiotics. Thus, all five CYP1 isoforms could be involved chemical effects, and perhaps in the resistance to AHR agonist toxicity observed in some Fundulus. Studies of the expression of the five CYP1 genes are under way in F. heteroclitus populations that have developed resistance to toxic effects of and a lack induction of CYP1A by halogenated AHR agonists (Elskus et al., 1999; Bello et al., 2001) .
In summary, we identified and cloned three new CYP1 genes, CYP1B1, CYP1C2 and CYP1D1, in the Atlantic killifish Fundulus heteroclitus, a vertebrate model used extensively in environmental toxicology studies. Substantial differences in the levels of expression for the five CYP1s were observed in the examined organs. We also showed that these CYP1s differ in the response to the potent AHR agonist PCB126. The organ-specific differences in basal levels, and in fold-induction by PCB126, suggest that regulatory mechanisms for the five CYP1 genes could differ. Finally, these new CYP1 family members increase the set of potential biomarkers of aquatic contamination in fish, and may aid in elucidating possible mechanisms of toxicity of AHR agonists, and the physiological roles of this important gene family.
This study was supported in part by NIH grants to JJS (the Superfund Basic Research Program 5P42ES007381 and R01ES015912) and MJJ (K99ES017044-01). JZ was a Guest Student at the Woods Hole Oceanographic Institution and was supported by a CAPES Ph.D. Fellowship and CNPq Ph.D. Sandwich Fellowship, Brazil. ACDB was recipient of the CNPQ Productivity Fellowship, Brazil. LW was supported by a Summer Student Fellowship at the Woods Hole Oceanographic Institution. Study sponsors had no involvement in the studies reported here or in the decision to submit this paper for publication.