Identification of NF-Kb in the Marine Fish Stenotomus Chrysops And

Chemico-Biological Interactions 126 (2000) 137–157 www.elsevier.com/locate/chembiont

Identiﬁcation of NF-kB in the marine ﬁsh Stenotomus chrysops and examination of its activation by aryl hydrocarbon receptor agonists

Jennifer J. Schlezinger a, Courtney E. Blickarz a, Koren K. Mann b, Stefan Doerre b, John J. Stegeman a,* a Biology Department, Redﬁeld 342, MS 32, Woods Hole Oceanographic Institution, Woods Hole, MA 02543, USA b Department of Microbiology, Boston Uni6ersity, Boston, MA 02118, USA Received 18 August 1999; received in revised form 23 February 2000; accepted 28 February 2000

Abstract

Members of the Rel family of proteins have been identified in Drosophila, an echinoderm, Xenopus, birds and mammals. Dimers of Rel proteins form the transcription factor nuclear factor kB (NF-kB) that rapidly activates genes encoding cytokines, cell surface receptors, cell adhesion molecules and acute phase proteins. Evidence suggests that xenobiotic compounds also may alter the activation of NF-kB. This study had a dual objective of identifying members of the Rel family and examining their activation by xenobiotic compounds in a marine fish model, scup (Stenotomus chrysops). A DNA-protein crosslinking technique demonstrated that liver, kidney and heart each had at least three nuclear proteins that showed specific binding to an NF-kB consensus sequence, with molecular weights suggesting that the proteins potentially corresponded to mouse p50, p65 (RelA) and c-rel. In addition, an :35kD NF-kB binding protein was evident in liver and kidney. The 50 kD protein was immunoprecipitated by mammalian p50-specific antibodies. The presence of Rel members in fish implied by those results was confirmed by RT-PCR cloning of a Rel homology domain

Abbre6iations: AhR, aryl hydrocarbon receptor; BP, benzo[a]pyrene; CYP, cytochrome P450; EMSA, electrophoretic mobility shift assay; DRE, dioxin response element; IP, immunoprecipitation; NF-kB, nuclear factor-kB; PeCB, 3,3%,4,4%,5-pentachlorobiphenyl; ROS, reactive oxygen species. * Corresponding author. Tel.:+1-508-2892320; fax: +1-508-4572169. E-mail address: [email protected] (J.J. Stegeman)

0009-2797/00/$ - see front matter © 2000 Published by Elsevier Science Ireland Ltd. All rights reserved. PII: S0009-2797(00)00161-7 138 J.J. Schlezinger et al. / Chemico-Biological Interactions 126 (2000) 137–157

(an apparent c-rel) from scup liver. NF-kB activation occurred in vehicle-treated fish, but this appeared to decrease over time. In fish treated with 0.01 or 1 mg 3,3%,4,4%,5-pentachlorobiphenyl per kg, NF-kB activation in liver did not decrease, and there was a 6–8-fold increase in activation 16–18 days following treatment. Treatment with 10 mg benzo[a]pyrene/kg had no effect on NF-kB-DNA binding, either at 3 or 6 days following treatment. The data show that the Rel family of proteins is present in fish, represented at least by a p50/105 homologue, and support a hypothesis that some aryl hydrocarbon receptor agonists can activate NF-kB in vivo. © 2000 Published by Elsevier Science Ireland Ltd. All rights reserved.

1. Introduction

Nuclear factor kB (NF-kB) is a family of dimeric transcription factors that can rapidly activate genes encoding cytokines, cell surface receptors, cell adhesion molecules and acute phase proteins [1]. NF-kB subunits are members of the Rel family of proteins, defined by the Rel homology domain, a 300-amino acid region which contains a DNA-binding region, a dimerization domain, a nuclear transloca- tion signal and a region of interaction with proteins of the IkB family [2,3]. Identified as the genes relish, dorsal, and dif in Drosophila [4–6], in mammals, the Rel family is represented by p65 (Rel A), RelB, c-rel, NF-kB1 (p52/p100) and NF-kB2 (p50/p105) (see [7]). Homologs of NF-kB1, p65, RelB and c-rel have been identified in Xenopus [8–11], homologs of NF-kB1, NF-kB2, p65 and c-rel have been sequenced from birds [12–16], and an NF-kB1 has been described in an echinoderm [17]. In this study, objectives included determining whether Rel family proteins are expressed in teleost fish and assessing their identity. Usually, NF-kB dimers are sequestered in the cytoplasm by an inhibitor, IkB. Upon stimulation by cytokines, mitogens, viral infection, UV irradiation or reactive oxygen species (ROS), IkB is phosphorylated, ubiquinated and digested by a proteosome, allowing NF-kB to translocate to the nucleus and bind DNA [18]. The ability of antioxidants to block NF-kB activation by potent activators such as tumor necrosis factor-a, interleukin-1, phorbol 12-myristate 13-acetate, lipopolysac- charide and okadaic acid [19] indicate that ROS may be second messengers that initiate degradation of IkB. Recent studies have shown that xenobiotic compounds also may activate NF-kB in vivo and in various cells in culture. In vivo, phenobarbital, ciprofibrate and carbon tetrachloride all have been shown to activate NF-kB binding in rat liver [20–22]. In vitro, 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD) was reported to transiently stimulate NF-kB binding in rat thymocytes [23] and mouse hepatoma cells [24], and 3,3%,4,4%-tetrachlorobiphenyl (TCB) stimulated NF-kB binding in porcine pulmonary arterial endothelial cells [25]. In addition, the polynuclear aromatic hydrocarbon 7,12-dimethylbenz[a]anthracene stimulates NF-kB binding in murine bone marrow stromal cells (Schlezinger et al., unpublished results). A sustained activation of NF-kB binding is seen following treatment of HepG2 cells with redox cycling xenobiotic quinones, including menadione and tert-butylhy- J.J. Schlezinger et al. / Chemico-Biological Interactions 126 (2000) 137–157 139 droquinone [26,27], suggesting that ROS may mediate the activation of NF-kBby some xenobiotics. The results with TCDD in cell cultures indicate that aryl hydrocarbon receptor (AhR) pathways may contribute to the activation of NF-kB. The AhR regulates the expression of drug metabolizing enzymes, including cytochromes P450 1As (CYP1A), and genes involved in cell growth control [28]. TCDD-mediated NF-kB activation in murine hepatoma cells has been reported to be AhR-dependent, ARNT-dependent and CYP1A1-dependent [24,29]. An antioxidant (N-acetylcys- teine) has been shown to block TCDD-mediated NF-kB activation [29], suggesting that ROS may be involved, which could be derived from CYP1A. CYP1A oxidizes polynuclear aromatic hydrocarbons to numerous metabolites including redox cycling quinones in mammals and fish [30–32], and planar halogenated aromatic hydrocarbons can stimulate ROS formation by uncoupling oxygen reduction and monooxygenation by CYP1A, established in studies with the fish scup [33,34]. Accordingly, in addition to investigating Rel protein expression in fish, we sought to determine whether NF-kB could be activated by AhR agonists in vivo, in the scup model. We investigated the effects of two AhR agonists, benzo[a]pyrene (BP) and the polychlorinated biphenyl 3,3%,4,4%,5-pentachlorobiphenyl (PeCB).

2. Materials and methods

2.1. Chemicals

3,3%,4,4%,5-Pentachlorobiphenyl (IUPACc 126) (PeCB) was purchased from Ultra Scientiﬁc (North Kingstown, RI). Poly [dI-dC], pNd6 and 5-bromo- 2%deoxyuridine were from Boehringer Mannheim (Indianapolis, IN). [g-32P]ATP and [a-32P]CTP were from Amersham (Arlington Heights, IL). All other chemical reagents were from Sigma Chemical Co. (St Louis, MO).

2.2. Animals and treatments

Scup (Stenotomus chrysops) were caught by trapping in Vineyard Sound, MA in August, 1995 and August, 1996, and the experiments were conducted in August, 1997 (1 year captive fish) and in June, 1998 (3 year captive fish). Fish were held in flow-through seawater tanks at 14°C and maintained on a diet of Purina Trout Chow, fed once weekly. Experimental animals (mixed sex) were gonadally undevel- oped and ranged in size from 39 to 327 g. In the first experiment, fish were injected intra-peritoneally with corn oil or suspensions of PeCB in corn oil at 0.01 or 1 mg/kg body weight (1 ml corn oil/kg body weight). For each dose group, three replicate tanks contained six fish each, and at each time point two fish were sampled from each tank. Untreated fish were killed by severing the spinal cord on day 0. Injected fish were killed 0, 3, 7, and 16 or 18 days following dosing. One of the five remaining fish in the 0.01 mg PeCB/kg dose group died between days 7 and 18, and three of the six remaining fish in the 1 mg PeCB/kg dose group died 140 J.J. Schlezinger et al. / Chemico-Biological Interactions 126 (2000) 137–157 between days 7 and 16. In the second experiment, fish were injected intra-peritoneally with corn oil or a suspension of BP in corn oil at 10 mg/kg body weight (1 ml corn oil/kg body weight). For each dose group, three tanks contained three fish each. At day 3 following injection, one fish from each tank was sampled, at a day 6 the remaining fish were sampled. Immediately following dissection, sections of liver, heart and kidney were frozen in liquid N2 and microsomes were prepared from portions of liver.

2.3. RNA isolation, RT-PCR and sequencing

Total RNA was prepared from liver using RNA-Stat 60 (Tel-Test Inc., Friendswoods, TX) by following the manufacturer’s instructions. RNA pellets were washed once in 4 M LiCl to remove glycogen, followed by two washes in 95% ethanol. Poly(A)+ RNA was isolated by one pass over a column of oligo(dT) cellulose (New England Biolabs, Beverly, MA) as described by Farrell [35]. Yields and puity were determined using a Shimadzu UV-2401PC spectrophotometer. First strand cDNA synthesis and subsequent amplification were performed using the Gene-AMP RNA-PCR kit (Perkin–Elmer, Foster City, CA), according to the manufacturer’s instructions. Poly(A)+ RNA (1 mg) was reverse transcribed with priming by random hexamers. Initial amplification was performed in the same tube with forward 5%-FRYVCEG-3% (5%-TT(TC)(CA)G(ACGT)-TA(TC)GT(ACGT)- TGTCGA(AG)GG-3%) and reverse 5%-DKVQKDD-3% (5%-C(AG)TC(AG)TC-(TC)- TT(TC)TG(ACGT)AC(TC)TT(AG)TC-3%) primers (at 0.15 mM each). PCR conditions were as follows: 105 s at 95°C followed by 35 cycles (15 s at 95°C and 30 s at 50°C). The last cycle was followed by extension for 7 min at 72°C. Aliquots (1 ml) of the original PCR reaction were re-amplified with forward 5%-AGSIPGE-3% (5%- GC(ACGT)GG(ACGT)TC-(ACGT)AT(ACT)CC(ACGT)GG(ACGT)GA(AG)-3%) and reverse 5%-CRVNKNCG-3% ((ACGT)-CC(AG)CA(AG)TT(CT)TT(AG)TT- (ACGT)AC(ACGT)C(TG)(AG)-CA-3%) primers (at 0.15 mM each). Conditions were as described above. PCR products were revealed by ethidium bromide staining after separation on a 1% agarose gel. The band of expected size (500 bp) was excised and purified (Geneclean II, Bio 101, La Jolla, CA). The PCR fragment was cloned into pGEM-T Easy vector (Promega, Madison, WI). Sequencing of both strands was performed using Sequenase (version 2.0, US Biochemical) with modifi- cations for direct sequencing as described by Bachmann et al. [36]. The sequence was confirmed by cycle-sequencing (SequiTherm long-read cycle sequencing kit. Epicentre Technologies, Madison, WI) using an automated sequencer (LI-COR, Inc. Lincoln, NE). Multiple alignments of the deduced amino acid sequences was performed using ClustalX [37]. Sites with alignment gaps were removed from the phylogenetic analyses. A Fitch-Margoliash distance tree was constructed using the programs PROTDIST and FITCH in the PHYLIP package (version 3.573c) created by J. Felsenstein. Protein distances were constructed from the Dayhoff et al. distances using global rearrangements and 1000 random sequence addition replicates. Boot- J.J. Schlezinger et al. / Chemico-Biological Interactions 126 (2000) 137–157 141 strapping was performed using 100 bootstrap replicates, global rearrangements, and 20 random sequence addition replicates per bootstrap replicate.

2.4. Microsome preparation and analysis of CYP1A

Microsomes were prepared by differential centrifugation [38]. Pellets were resus- pended in buffer (50 mM Tris, pH 7.4, 1 mM dithiothrietol, 1 mM EDTA,

20% glycerol) and frozen in liquid N2 until use. Protein content was determined using the bicinchoninic acid method, with bovine serum albumin standards. Micro- somal proteins (0.1–0.4 mg/slot) were applied to a 0.05 mm nitrocellulose membrane using a slot blot apparatus. The primary antibody was the mouse monoclonal antibody 1-12-3 [39]. The secondary antibody was a horseradish peroxidase-linked sheep anti-mouse IgG (Amersham). The immunoreactive bands were visualized using enhanced chemiluminescence (Amersham) and quantiﬁed by video image analysis (NIH Image 1.60b5) by comparison to scup standards of known concentration.

2.5. Gel-shift analysis of NF-sB acti6ation

Crude nuclear extracts were prepared from tissues (100–300 mg) by the method of Deryckere and Gannon [40]. Protein content was determined as above. Nuclear extracts were stored at −80°C. For analysis of NF-kB activation, the double- stranded oligonucleotide 5%-AGTTGAGGGGACTTTCCCAGGC-3% (Promega) representing the consensus binding site was used. The DNA probe was end-labeled using T4 polynucleotide kinase (Promega) and [g-32P]ATP and was purified using a Centrispin-20 column (Princeton Separations, Adelphia, NJ). The electrophoretic mobility shift assay (EMSA) was performed as follows: the 32P-labeled DNA (:0.5 ng, 50 000 cpm) and 2 mg of nuclear extract were combined with buffer (final concentrations: 10 mM Tris–HCl pH 7.5, 1 mM EDTA, 40 mM KCl, 0.5 mM m MgCl2, 1 mM DTT, 10% glycerol) and poly dI-dC (1 g) in a final volume of 20 ml. The mixture was incubated at room temperature for 30 min. Polyacrylamide gels (6%) were pre-run at 80 V for 1 h. Mixtures were electrophoresed at 80 V for 1.5 h in 0.25×TBE (final concentrations: 22 mM TRIS–base pH 8, 22 mM boric acid, 0.4 mM EDTA). The gels were dried and exposed to film. Shifted bands were quantified by video image analysis (NIH Image 1.60b5). Specificity of shifted bands was determined by including 50–100×cold oligonucleotides containing NF-kB consensus sequences (IL2Ra palindromic variant: 5%-CGGCAGGGGAATTCCC- CTCTCC-3%; HIV-LTR: 5%-CCGCTGGGGACTTTCCAGGC-3%) or mutated ver- sions of those sequences (IL2RakB mutant: 5%-CAACGGCAGATCTATCTCC- CTCTCCTT-3%; HIV mutant: 5%-CCGCTGATCACTT-TCCAGGC-3%) [41].

2.6. Identiﬁcation of speciﬁc subunits

To identify specific NF-kB/Rel subunits, DNA-protein crosslinking studies were used [41,42]. A photoreactive 32P-labeled oligonucleotide was prepared by annealing 142 J.J. Schlezinger et al. / Chemico-Biological Interactions 126 (2000) 137–157 the oligonucleotide containing the IL2Ra palindromic variant (5%-CAACG- GCAGGGGAATTCCCCTCTCCTT-3%) with the primer (5%-AAGGAGAGGG-3%) and using the Klenow fragment of DNA polymerase (Promega) [42]. The binding reaction was performed as follows: 5–20 mg of nuclear extract were combined with buffer (Final concentrations: 20 mM HEPES pH 7.9, 1 mM DTT, 1 mM EDTA, m m m 5% glycerol), poly dI-dC (1 g), BSA (5 g), pdN6 (1 g) in a final volume of 20 ml. The mixture was incubated at room temperature for 5 min. The 32P-labeled DNA (:2.5–5 ng, 200 000-500 000 cpm) was added, and the mixture was incubated for 15 min at room temperature. In reaction mixtures containing competitors, the competitors (100-fold excess) were added 15 min prior to the addition of the probe. The mixture was then incubated for 15 min under UV. At this point, some samples were boiled with sample treatment buffer and frozen at −20°C. Some reaction mixtures were immunoprecipitated with antibodies to specific NF-kB/Rel subunits (a-p50: sc-114, human NLS; a-p52: sc-848, human full length; a-p65: sc-109, human amino-terminal region; a-crel: sc-272, human amino-terminal region; Santa Cruz Biotechnology). Sample treatment buffer was added to the immunopre- cipitates, and they were boiled for 5 min. All samples then were loaded on a 10% SDS polyacrylamide gel and electrophoresed at 10–14 mAmps. The portion of the gel containing free probe was removed. The gel was fixed in MeOH (5%)/acetic acid (5%) for 30 min, dried and exposed to film.

2.7. Statistics

Statistics were calculated using Microsoft Excel (Microsoft, Inc., Redmond, WA) and SuperAnova for Macintosh (Abacus Concepts, Inc., Berkeley, CA). The Student’ t-test was used to determine differences from control values in the BP experiment. Nested, one-factor ANOVAs with the Tukey–Kramer multiple com- parisons test were used to analyze differences between treatment groups within sampling days for the PeCB experiment.

3. Results

3.1. Binding of an NF-sB consensus sequence to teleost nuclear proteins

One goal of this study was to determine the presence of NF-kB-like proteins in teleost fish and to characterize these proteins using biochemical and cloning approaches. Using an EMSA, we determined that nuclear proteins from liver, kidney and heart of naive (untreated) scup bound an NF-kB consensus sequence (Fig. 1). In order to confirm the specificity of binding and that Rel protein were responsible for the band-shift of the NF-kB-binding oligonucleotide, we performed a series of competition, crosslinking and immunoprecipitation (IP) studies using nuclear protein from the various tissues. Adding a 50–100 fold excess of unlabeled oligonucleotides containing various NF-kB binding sequences eliminated the band- shift by liver nuclear proteins (Fig. 2). The homologous competitors included the J.J. Schlezinger et al. / Chemico-Biological Interactions 126 (2000) 137–157 143

IL-2Ra palindromic variant binding site used in the crosslinking studies (see below) and the HIV-LTR binding site found to be occupied by NF-kB following AhR agonist treatment in Hepa cells [29]. An oligonucleotide containing a single mutated site (5%-CAACGGCAGGGCTATCTCCCTCTCCTT-3%) also was able to compete the band-shift (data not shown), while a 50–100 fold excess of unlabeled oligonu-

Fig. 1. Nuclear proteins from scup liver, kidney and heart bound an NF-kB consensus sequence. Nuclear extracts were prepared from naive ﬁsh and analyzed by EMSA for 32P-labeled NF-kB oligonucleotide binding. Reactions contained 2, 1 and 5 mg of nuclear protein from liver, kidney and heart, respectively. Proteins from individual ﬁsh were resolved in each lane. Lane 1: free probe. Lanes 2–4: liver. Lanes 5–7: kidney. Lanes 8–10: heart.

Fig. 2. Scup liver nuclear proteins specifically bound an NF-kB consensus sequence. The specificity of NF-kB binding activity was tested by the addition of excess of unlabeled NF-kB binding oligonucleotides or mutated NF-kB binding oligonucleotides. Nuclear extracts were prepared from naive fish and analyzed by EMSA for 32P-labeled NF-kB oligonucleotide binding. Reactions contained 2 mg of nuclear protein. Lane 1: no competition. Lane 2 and 3: competition with the palindromic variant of the IL2Ra kB site (5%-CGGCAGGGGAATTCCCCTCTCC-3%) at 50 and 100-fold excess. Lanes 4 and 5: competition with a mutated IL2RakB site (5%-CAACGGCAGATCTATCTCCCTCTCCTT-3%)at50and 100-fold excess. Lanes 6 and 7: competition with an HIV kB site (5%-CCGCTGGGGACTTTCCAGGC- 3%) at 50 and 100-fold excess. Lanes 8 and 9: competition with a mutated HIV kB site (5%-CCGCTGAT- CACTTTC-CAGGC-3%) at 50 and 100-fold excess. 144 J.J. Schlezinger et al. / Chemico-Biological Interactions 126 (2000) 137–157

Fig. 3. Characterization of scup liver nuclear proteins that bind an NF-kB consensus sequence. The identity of the subunits binding to the NF-kB site was analyzed by protein crosslinking (A) and IP (B). Liver nuclear extracts prepared from naive scup or from WEHI 231 cells were incubated with a 5-bromo-2%-deoxyuridine-substituted 32P-labeled NF-kB oligonucleotide probe and crosslinked by UV irradiation. The resultant labeled DNA-protein complexes also were immunoprecipitated with antibodies to specific NF-kB subunits. All were resolved on a 10% polyacrylamide gel. A). Lane 1: molecular weight standards corresponding to 200, 97, 68, 43 and 29 kD are indicated by black bars. Lanes 2–4: crosslinking of nuclear proteins from 3 individual fish. Lane 5: competition with the palindromic variant of the IL2RakB site. Lane 6: competition with a mutated IL2RakB site. B). Data shown are from different exposures of the same gel. Lane 1: WEHI 231 nuclear proteins crosslinked and IPed with anti-p50 antibody. Lane 2: WEHI 231 nuclear proteins crosslinked and IPed with anti-p65 antibody. Lane 3: WEHI 231 nuclear proteins crosslinked and IPed with anti-c-rel antibody. Lane 4: scup liver nuclear proteins crosslinked and IPed with anti-p50 antibody. cleotides containing three mutated sites did not inhibit the band-shift (Fig. 2). Liver nuclear extracts from vehicle or AhR-agonist treated fish did not bind to the consensus binding sequences of AP-1, another redox sensitive transcription factor [43], or Oct1, a ubiquitous mammalian transcription factor [44] (data not shown). Using a protein crosslinking technique, it was shown that the scup hepatic nuclear proteins that bound the NF-kB consensus had electrophoretic migrations that were similar to mouse p50, p65 and c-rel (Fig. 3A). In addition, a smaller protein of :35 kD bound the NF-kB site (Fig. 3A). The :50 kD protein could be immunoprecipitated with a mammalian p50-specific antibody (Fig. 3B). Anti- bodies to mammalian p52, p65, c-Rel or RelB did not recognize the other proteins under the conditions of assay (data not shown). EMSA analysis with competitors showed that nuclear proteins from scup kidney (Fig. 4A) and heart (Fig. 5A) also specifically bound the NF-kB consensus J.J. Schlezinger et al. / Chemico-Biological Interactions 126 (2000) 137–157 145 sequence. Similar to liver, at least three nuclear proteins from scup heart and kidney bound the NF-kB consensus sequence (Fig. 4B and 5B). As in liver, an :35kD kB site binding protein was evident in kidney. Also as with the liver proteins, one of the proteins (:50 kD) in heart and kidney was recognized by a mammalian p50-specific antibody (Fig. 4B and Fig. 5B).

3.2. Identiﬁcation of a teleost Rel homology domain

The presence of Rel family proteins was conﬁrmed further by cloning of a Rel homology domain from scup liver. Nested primers were designed to target the Rel homology domain of Xenopus Rel family proteins. Reverse transcription followed by PCR resulted in the production of a band of the expected size (500 bp). This band was cloned and sequenced (Fig. 6). The BLAST algorithm [45] was used to align the scup sequence with other known sequences in the Genbank data base. A phylogenetic analysis showed that the scup Rel sequence was in the c-rel cluster and that this putative scup c-rel homologue grouped most closely with Xenopus xRel2 (Fig. 7).

Fig. 4. Characterization of scup kidney nuclear proteins that bind an NF-kB consensus sequence. (A) The speciﬁcity of NF-kB binding activity was tested by the addition of excess of unlabeled NF-kB binding oligonucleotides or mutated NF-kB binding oligonucleotides as in Fig. 3. Lane 1: free probe. Lane 2: no competition. Lane 3: competition with the palindromic variant of the IL2RakB site at 50-fold excess. Lane 4: competition with a mutated IL2RakB site at 50-fold excess. Lane 5: competition with an HIV kB site at 50-fold excess. Lane 6: competition with a mutated HIV kB site at 50-fold excess. (B) The identity of the subunits binding to the NF-kB site was analyzed by protein crosslinking and IP as in Fig. 3. Lane 1: molecular weight standards corresponding to 203, 135, 86, 42 and 33 kD are indicated by black bars. Lanes 2–4: crosslinking of nuclear proteins from 3 individual ﬁsh. Lane 5: IP with anti-p50 antibody. 146 J.J. Schlezinger et al. / Chemico-Biological Interactions 126 (2000) 137–157

Fig. 5. Characterization of scup heart nuclear proteins that bind an NF-kB consensus sequence. (A) The speciﬁcity of NF-kB binding activity was tested by the addition of excess of unlabeled NF-kB binding oligonucleotides or mutated NF-kB binding oligonucleotides as in Fig. 3. Each reaction contained 5 ug of nuclear protein. Lane 1: free probe. Lane 2: no competition. Lane 3: competition with the palindromic variant of the IL2RakB site at 50-fold excess. Lane 4: competition with a mutated IL2RakB site at 50-fold excess. Lane 5: competition with an HIV kB site at 50-fold excess. Lane 6: competition with a mutated HIV kB site at 50-fold excess. (B) The identity of the subunits binding to the kB site was analyzed by protein crosslinking and IP as in Fig. 3. Lane 1: molecular weight standards corresponding to 203, 135, 86, 42 and 33 kD are indicated by black bars. Lanes 2–4: crosslinking of nuclear proteins from 3 individual ﬁsh. Lane 5: IP with anti-p50 antibody.

3.3. Effect of AhR agonist treatment on NF-sB acti6ation

Liver from fish treated with either of two AhR agonists, PeCB or BP, were examined for activation of the AhR and of NF-kB-DNA binding. As expected, PeCB activated the AhR, as indicated by induction of CYP1A protein (Fig. 8). As has been seen with another non-ortho substituted polychlorinated biphenyl 3,3%,4,4%- tetrachlorobiphenyl [46], the high dose of PeCB (1 mg/kg) suppressed the level of CYP1A induction, probably the result of post-transcriptional inactivation of the protein (Fig. 8). Hepatic nuclear proteins from both vehicle-treated and PeCB-treated scup bound an NF-kB consensus sequence (Fig. 9A). Corn oil treatment alone may have contributed to the activation of NF-kB; the NF-kB band density in naive fish was 18.997.8 integrated density/mg nuclear protein and that in fish 3 days post-corn oil treatment was 26.3918.3 integrated density/mg nuclear protein. However, NF-kB activation appeared to decrease over time in the vehicle-treated controls (Fig. 9B). In contrast, NF-kB activation was significantly elevated in PeCB-treated fish over time until 16–18 days post-treatment, at which time there was a 6-fold greater activation with the low dose and an 8-fold greater activation with the high dose of J.J. Schlezinger et al. / Chemico-Biological Interactions 126 (2000) 137–157 147

PeCB that in control ﬁsh (Fig. 9B). In a second experiment in which NF-kB activation was assessed only at one time point (14 days), a slight increase in activation occurred following a low dose (0.01 mg/kg) PeCB exposure. However, the variability was very large (data not shown), and the differences were not signiﬁcant. BP activated the AhR, as shown by the strong induction of CYP1A protein at days 3 (Control: 15911 pmol CYP1A/mg, BP: 6509155 pmol/mg) and 6 (Con- trol: 15912 pmol/mg, BP: 5959206 pmol/mg). Hepatic nuclear proteins from both control and BP-treated scup bound an NF-kB consensus sequence. NF-kB activation decreased over time in the vehicle-treated controls; however, no change in NF-kB activation occurred following BP treatment (Fig. 10).

Fig. 6. Nucleotide (A) and deduced amino acid (B) sequences of a Rel homology domain from a teleost. The amino acid sequence was aligned using the program ClustalX. Genbank accession numbers are as follows: Xenopus Xrel2, Z49252, [11]; turkey c-rel, K02455, [12]; chicken c-rel, X52193, [13]; mouse c-rel, X15842 [63]; human c-rel, X75042 [64]. * The same amino acid is present in all sequences. 148 J.J. Schlezinger et al. / Chemico-Biological Interactions 126 (2000) 137–157

Fig. 7. Phylogenetic analysis of a Rel protein sequence from scup. The best Fitch–Margoliash tree was found by FITCH (see Section 2). Only the portion of sequence shown in Fig. 6 was used for the analysis. Sites with alignment gaps were removed from the analysis. The single best tree is shown, with bootstrap numbers superimposed (bold number next to branch point). Branch lengths represent evolutionary change. Genbank accession numbers are as follows: Dif, L29015 [5]; Dorsal, M23702 [6]; Relish, U62005 [4]; urchin Rel, AF064258 [17]; Xenopus Xrel1, M60785 [8]; Xenopus RelB, D63332 [9]; Xenopus p100, AB002629 [10]; Xenopus Xrel2, Z49252 [11]; turkey c-rel, K02455 [12]; chicken c-rel, X52193 [13]; chicken p105, M86930 [14]; chicken RelA, D13721 [15]; chicken p100, U00111 [16]; mouse p105, M57999 [65]; mouse RelA, M61909 [66] mouse RelB, M83380 [67]; mouse c-rel, X15842 [63]; human p105, M55643 [68]; human p105, S76638 [69] human c-rel, X75042 [64]; human RelA, M6239 [70]; human RelB, M83221 [71].

4. Discussion

The transcription factor, NF-kB, is comprised of dimers of members of the Rel protein family. These studies pursued two related objectives. First, we demonstrated the presence of Rel family members and NF-kB binding to DNA in teleost fish, a previously unstudied group. Second, we determined that NF-kB binding to DNA could be activated in scup liver following in vivo treatment with a planar halogenated aromatic hydrocarbon AhR agonist. Likely function of NF-kB in gene regulation in fish is indicated by the identification of NF-kB binding sites in the J.J. Schlezinger et al. / Chemico-Biological Interactions 126 (2000) 137–157 149 promoter of mitogen-activated protein kinase kinase 1 of carp [47] and also in a killifish cytochrome P450 2 gene, CYP2P1 (Oleksiak and Stegeman, unpublished results).

4.1. Identiﬁcation of Rel proteins in a teleost

Considering the important function of Rel family proteins during development and in the immune system [5,6,11,48], these proteins presumably should occur broadly and perhaps in all animals. However, there remains a question of how conserved are the structure and function of these proteins. Rel proteins are well characterized in mammals (see Ref. [7]) and have been identified in birds [17], an amphibian [8–10] and an echinoderm [17], but there is a lack of knowledge on Rel proteins in fish. The molecular and biochemical evidence provided here indicates the presence of multiple Rel family proteins in the teleost fish, scup.

1. A sequence containing a Rel homology domain was cloned from scup liver. Alignment analyses suggest that all Rel/NF-kB and IkB proteins belong to a superfamily of genes whose precursor contained both a Rel homology domain and an ankyrin repeat area [7]. Molecular phylogenetic analysis indicates that the scup sequence is a c-rel homologue. 2. Nuclear proteins that speciﬁcally bound to an oligonucleotide containing an NF-kB consensus binding sequence were identiﬁed in multiple organs of scup. The fact that a single point mutation in the IL-2Ra NF-kB binding site did not

Fig. 8. Effect of PeCB on hepatic microsomal CYP1A content. Scup were injected intra-peritoneally with corn oil, 0.01 mg PeCB/kg or 1 mg PeCB/kg and killed 7, 16, or 18 days post-injection. Liver microsomal fractions were applied to a nitrocellulose membrane with a slot blot apparatus, immunoblot- ted with MAb 1-12-3 against scup CYP1A and visualized using enhanced chemiluminescence. Data represent the means9S.D. of measurements on six fish per treatment group except for the 0.01 mg/kg dose on Day 18 (n=4) and the 1 mg/kg dose on Day 16 (n=3). c Significantly different from control (PB0.01, Tukey–Kramer). * Significantly different from control and 1 mg/kg dose groups (PB0.01, Tukey–Kramer). 150 J.J. Schlezinger et al. / Chemico-Biological Interactions 126 (2000) 137–157

Fig. 9. Quantification of NF-kB activation in liver of control and PeCB-treated scup. (A) EMSA of hepatic nuclear proteins from fish treated with corn oil or PeCB for 16–18 days. (B) Quantification of NF-kB-DNA binding in hepatic nuclear proteins from corn oil or PeCB treated fish. Treatment of fish was as described in Fig. 8. Nuclear extracts were prepared from livers and analyzed by EMSA for 32P-labeled NF-kB oligonucleotide binding. Nuclear extracts were diluted with water and combined with aliquots of the same reaction mixture containing the labeled probe. All reactions proceeded at the same time and were loaded and run on gels at the same time. All gels were exposed to the same piece of film. The first 3 lanes of the EMSA figure were run on one gel while the remaining lanes were run on a second gel. Band intensity was quantified using video image analysis. Data represent the means9S.D. of measurements on 6 fish per treatment group except for the 0.01 mg/kg dose on Day 18 (n=4) and the 1mg/kg dose on Day 16 (n=3). c Significantly different from control (PB0.01, Tukey–Kramer). J.J. Schlezinger et al. / Chemico-Biological Interactions 126 (2000) 137–157 151

Fig. 10. Quantification of binding of an NF-kB consensus sequence to liver nuclear proteins from control and BP-treated scup. Nuclear extracts prepared from livers from fish treated with corn oil or 10 mg/kg BP and killed on days 3 and 6 were analyzed by EMSA for 32P-labeled NF-kB oligonucleotide binding. Band intensity was quantified using video image analysis. Data represent mean9S.D. of three fish per treatment group on day 3 and six fish per treatment group on day 6. * Significantly different from day 3 (PB0.03, Student’s t-test).

abrogate binding and that binding could be competed by NF-kB sites of varying sequence suggests that the binding of NF-kB is somewhat promiscuous in ﬁsh as it is in mammals [49]. 3. Protein/DNA crosslnking demonstrated three proteins with molecular weights similar to murine p50, p65 and c-rel. 4. The :50 kD binding protein was recognized by a mammalian p50-speciﬁc antibody. Among the Rel proteins, p50/105 and p52/100 genes have been found to evolve at the slowest rate among the Rel proteins [7], thus the mammalian p50 antibody was most likely to recognize a teleost protein. The fact that antibodies to p65 and c-rel did not recognize the larger molecular weight proteins could suggest that these proteins are not NF-kB subunits, or more likely, suggests that enough difference exists between these proteins and the mammalian homologues to obscure cross-the epitope recognized by the antibody.

4.2. Acti6ation of NF-sB DNA binding by an AhR agonist

Results from this study indicate that at least one AhR agonist, PeCB, can activate NF-kB-DNA binding in vivo in teleost liver. The in vivo results with PeCB are consistent with results from experiments in cell culture. Those studies suggested that the activation of NF-kB by an AhR agonist requires the AhR [29]. Liver is an appropriate site to analyze NF-kB activation, as NF-kB is a mediator of hepatocyte proliferation; NF-kB is rapidly activated following partial hepatectomy [50], and livers of p65 knockout mice undergo massive apoptosis early in gestation [51]. Evaluating the activation of NF-kB binding to DNA by PeCB was, however, confounded somewhat by the apparent activation of NF-kB in fish injected with 152 J.J. Schlezinger et al. / Chemico-Biological Interactions 126 (2000) 137–157 corn oil alone. It was reported recently that corn oil activated NF-kB in Kupffer- cells [52], and it is possible that corn oil had a similar effect in this study. Nevertheless, there was an apparent dose-dependence in the activation by PeCB. In fish given the low dose and in those sampled at intermediate times, this activation was relative; the activation in the controls decreased over time, while in the treated fish, it did not. In fish given the high dose of PeCB there was a significant increase in activation, but that occurred only more than 2 weeks after treatment. While PeCB appeared to activate NF-kB, another AhR agonist, BP, did not. One important difference between the sampling regimes in each experiment was that BP-treated fish were sampled only at early time points (i.e. 3–6 days); the increase in NF-kB binding did not occur in PeCB-treated fish until day 16. The TCDD-mediated activation of expression of a cat reporter containing an NF-kB binding sequence in Hepa cells occurred 24 h post-treatment, suggesting that the activation required the AhR [29]. However, that activation could be blocked by anti-oxidants [29]. Thus, ROS production may be a potential mechanism by which AhR agonists activate NF-kB. Other compounds appear to activate NF-kB via ROS. Redox cycling quinones activate NF-kB in HepG2 cells, apparently as a protective mechanism [26,27]. Phenobarbital and ciprofibrate have been shown to activate NF-kB-DNA binding in rat liver [21,22]. Phenobarbital induces CYP2B1/2B2, which was suggested to produce ROS as a metabolic by-product. Ciprofibrate, a peroxisome proliferator, induces enzymes in the peroxisomal b-oxidation pathway which also produce ROS as by-products. The TCDD-mediated activation of NF-kB binding in Hepa cells also required a functional CYP1A1 [29], which may be a link to ROS generation. Planar halogenated aromatic hydrocarbons, including the coplanar polychlorinated biphenyl PeCB used here, stimulate ROS production by uncoupling the CYP1A catalytic cycle [34,53] which is consistent with the requirement for functional CYP1A in activation of NF-kB by TCDD [29]. BP is metabolized to redox cycling quinones in fish and mammals [31]. The apparent lack of NF-kB activation in BP-treated fish could result if the BP metabolites in vivo were rapidly conjugated and/or excreted. It does not appear that AhR agonists activate NF-kB simply by activating the AhR. In scup, pronounced increases in AhR-mediated gene expression (i.e. CYP1A) were seen within 7 days. The reason for the delay in NF-kB activation is unknown. However, assuming that CYP1A generation of ROS is involved, then the timing of competing events could contribute to a delay in NF-kB activation. First, analysis of PeCB-treated fish indicated that antioxidant enzyme activities were increased and peaked at 3–7 days, while the level of CYP1A remains elevated for at least a week longer (Schlezinger and Stegeman, unpublished data). There- fore, any ROS produced may not overwhelm the cell until after the decrease in anti-oxidant enzyme activity. Second, at high doses, PeCB inactivates CYP1A in vitro [34] and in vivo (Schlezinger and Stegeman, unpublished data). Production of CYP1A protein may be overwhelmed by inactivation at early time points, there- fore ROS production is low. As the PeCB concentration decreases in the liver, a point may be reached when the PeCB concentration is high enough to stimulate J.J. Schlezinger et al. / Chemico-Biological Interactions 126 (2000) 137–157 153

ROS production but low enough not to significantly inactivate CYP1A. A relief of inactivation is seen with another planar PCB, 3,3%,4,4%-tetrachlorobiphenyl, at approximately 14 days post-injection [46]. In addition to the suggested involvement of CYP1A, the AhR and NF-kB also may interact directly. An overlapping DRE/NF-kB site has been identified in the 3% a-hs4 enhancer region of the murine immunoglobulin heavy chain locus [54]. This site is always bound by an NF-kB dimer containing p50 but also is bound by the AhR in B cells treated with TCDD. Further, p65 can be co-immunoprecipitated with AhR in cytosol or nuclear fractions of TCDD-treated Hepa cells and TCDD-treated COS cells that are cotransfected with AhR and p65 [55]. Tian et al. suggested that TCDD down-regulates NF-kB mediated cat reporter activity in Hepa cells [55]. However, the only construct tested in the analysis contained five consecutive NF-kB sites; how this might affect the results is not known. The activation of NF-kB following treatment with AhR agonists in vitro and in vivo suggests that the AhR can influence the regulation of NF-kB. How would these two transcription factors be linked? Activation of NF-kB may protect against the adverse effects of xenobiotic metabolism by regulating genes involved in the antioxidant response. Following NF-kB activation, induction of Mn-superoxide dismutase [56], NADPH-oxidoreductase [57] inducible nitric oxide synthase [58], an ferritin H [59] has been reported. NF-kB activation was suggested to be the mechanism by which HepG2 cells are protected from oxidative stress-induced cytotoxicity [27]. In addition, an NF-kB consensus binding sequence has been identified in the promoter of the P-glycoprotein gene [60]. P-glycoprotein can be induced by TCDD [61], and xenobiotics may be removed from cells via the P-glycoprotein transporter [62]. The mechanisms of the regulation of NF-kBby the AhR have just begun to be elucidated. Future experiments should address, first, the possibility that the AhR may directly regulate the transcription of at least one NF-kB subunit; the p50/105 promoter contains overlapping DREs (−527– −511, CTGCGTGCGCGCGTGTGTCC). Second, these two transcription factors also appear to interact during DNA binding, but the binding site(s) occupied by the AhR and NF-kB during this interaction remain to be determined. Examining the nature of the regulation of NF-kB by the AhR and AhR agonists may be important in determining the role of NF-kB in mediating xenobiotic effects.

Acknowledgements

This research was supported in part by EPA grant R827102-01, NIH grant P42-ES07381, and by a WHOI Mellon Award to J. Stegeman and by the Lyons Fellowship at MIT to J. Schlezinger. We gratefully acknowledge Bruce Woodin for technical assistance and Dr Andrew McArthur for assistance with the phylogenetic analysis. This is Contribution Number 10126 of the Woods Hole Oceano- graphic Institution. 154 J.J. Schlezinger et al. / Chemico-Biological Interactions 126 (2000) 137–157

References

[1] P.A. Baeuerle, The inducible transcription activator NF-kB: regulation by distinct protein subunits, Biochim. Biophys. Acta 1072 (1991) 63–80. [2] I.M. Verma, J.K. Stevenson, E.M. Schwartz, D. Van Antwerp, S. Miyamoto, Rel/NF-kB/IkB family: intimate tales of association and dissociation, Genes Dev. 9 (1995) 2723–2735. [3] S. Miyamoto, I.M. Verma, Rel/NF-kappa B/I kappa B story, Adv. Cancer Res. 66 (1995) 255–292. [4] M.S. Dushay, B. Asling, D. Hultmark, Origins of immunity: relish, a compound Rel-like gene in the antibacterial defense of Drosophila, Proc. Natl. Acad. Sci. USA 93 (1996) 10343–10347. [5] Y.T. Ip, M. Reach, Y. Engstrom, et al., Dif, a dorsal-related gene that mediates an immune response in Drosophila, Cell 75 (1993) 753–763. [6] R. Steward, Dorsal, an embryonic polarity gene in Drosophila, is homologous to the vertebrate proto-oncogene, c-rel, Science 238 (1987) 692–694. [7] C. Huguet, P. Crepieux, V. Laudet, Rel/NF-kB transcription factors and IkB inhibitors: evolution from a unique common ancestor, Oncogene 15 (1997) 2965–2974. [8] K.R. Kao, N.D. Hopwood, Expression of a mRNA related to c-rel and dorsal in early Xenopus lae6is embryos, Proc. Natl. Acad. Sci. 88 (1991) 2697–2701. [9] K. Suzuki, T. Yamamoto, J. Inoue, Molecular cloning of cDNA encoding the Xenopus homolog of mammalian RelB, Nuc. Acids Res. 23 (1995) 4664–4669. [10] K. Suzuki, J. Tsuchida, T. Yamamoto, J. Inoue, Identification and expression of the Xenopus homolog of mammalian p100-NFkB2, Gene 206 (1998) 1–9. [11] D. Tannahill, F.C. Wardle, Control of axis formation in Xenopus by the NF-kB-IkB system, Int. J. Dev. Biol. 39 (1995) 549–558. [12] K.C. Wilhelmsen, K. Eggleton, H.M. Temin, Nucleic acid sequences of the oncogene v-rel in reticuloendotheliosis virus strain T and its cellular homolog, the proto-oncogene c-rel, J. Virol. 52 (1984) 172–182. [13] A.J. Capobianco, D.L. Simmons, T.D. Gilmore, Cloning and expression of a chicken c-rel cDNA: unlike p59v-rel, p68c-rel is a cytoplasmic protein in chicken embryo fibroblasts, Oncogene 5 (1990) 257–265. [14] A.J. Capobianco, D. Chang, G. Mosialos, T.D. Gilmore, p105, the NF-kB p50 precursor protein, is one of the cellular proteins complexed with the v-Rel oncoprotein in transformed chicken spleen cells, J. Virol. 66 (1992) 3758–3767. [15] T. Ikeda, K. Honjo, Y. Hirota, T. Onodera, Isolation of the chicken NF-kB p65 subunit-encoding cDNA and characterization of its products, Gene 133 (1993) 237–242. [16] S. Sif, T.D. Gilmore, NF-kB p100 is one of the high-molecular-weight proteins complexed with the v-Rel oncoprotein in transformed chicken spleen cells, J. Virol. 67 (1993) 7612–7617. [17] Z. Pancer, J. Rast, E.H. Davidson, Origins of immunity: transcription factors and homologues of effector genes in sea urchin coelomocytes, Immunogenetics 49 (1999) 773–786. [18] A.S.J. Baldwin, The NF-kB and IkB proteins: new discoveries and insights, Annu. Rev. Immunol. 14 (1996) 649–681. [19] R. Schreck, P. Rieber, P.A. Baeuerle, Reactive oxygen intermediates as apparently widely used messengers in the activation of the NF-kB transcription factor and HIV-1, EMBO J. 10 (1991) 2247–2258. [20] A. Gruebele, K. Zawaski, D. Kaplan, R.F. Novak, Cytochrome P4502E1- and cytochrome P4502B1/2B2-catalyzed carbon tetrachloride metabolism: effects on signal transduction as demonstrated by altered immediate-early (c-Fos and c-Jun) gene expression and nuclear AP-1 and NF-kB transcription factor levels, Drug Metab. Dispo. 24 (1996) 15–22. [21] X. Li, L.K. Leung, B.T. Spear, H.P. Glauert, Activation of hepatic NF-kB by phenobarbital in rats, Biochem. Biophys. Res. Commun. 229 (1996) 982–989. [22] Y. Li, L.K. Leung, H.P. Glauert, B.T. Spear, Treatment of rats with the peroxisome proliferator ciprofibrate results in increased liver NF-kB activity, Carcinogenesis 17 (1996) 2305–2309. [23] M.J. Olnes, M. Verma, R.N. Kurl, 2,3,7,8-Tetrachlorodibenzo-p-dioxin-mediated gene expression in the immature rat thymus, Exp. Clin. Immunogenet. 11 (1994) 102–109. J.J. Schlezinger et al. / Chemico-Biological Interactions 126 (2000) 137–157 155

[24] S.J. Barnes, C.Y. Chang, H. Zhu, et al., Activation of redox-controlled transcription factors by TCDD, Toxicol. Sci. 48 (1999) 63. [25] B. Hennig, R. Slim, M. Toborek, L.W. Robertson, Linoleic acid amplifies polychlorinated biphenyl- mediated dysfunction of endothelial cells, J. Biochem. Mol. Toxicol. 13 (1999) 83–91. [26] R. Pinkus, L.M. Weiner, V. Daniel, Role of oxidants and antioxidants in the induction of AP-1, NF-kB, and glutathione S-transferase gene expression, J. Biol. Chem. 271 (1996) 13422–13429. [27] Q. Chen, A.I. Cederbaum, Menadione cytotoxicity to Hep G2 cells and protection by activation of nuclear factor-kB, Mol. Pharmacol. 52 (1997) 648–657. [28] O. Hankinson, The aryl hydrocarbon receptor complex, Ann. Rev. Pharmacol. Toxicol. 35 (1995) 307–340. [29] Y. Yao, A. Hoffer, C.-Y. Chang, A. Puga, Dioxin activates HIV-1 gene expression by an oxidative stress pathway requiring a functional cytochrome P450 CYP1A1 enzyme, Environ. Health Perspect. 103 (1995) 366–371. [30] R.J. Lorentzen, P.O.P. Ts’o, Benzo[a]pyrenedione/benzo[a]pyrenediol oxidation-reduction couples and the generation of reactive reduced molecular oxygen, Biochemistry 16 (1977) 1467–1473. [31] A.V. Klotz, J.J. Stegeman, C. Walsh, An aryl hydrocarbon hydroxylating hepatic cytochrome P-450 from the marine fish Stenotomus chrysops, Arch. Biochem. Biophys. 226 (1983) 578–592. [32] J.J. Stegeman, Polynuclear aromatic hydrocarbons and their metabolism in the marine environ- ment, in: H.V. Gelboin, P.O.P. Ts’o (Eds.), Polycyclic Hydrocarbons and Cancer, Academic Press, New York, 1981, pp. 1–60. [33] J.J. Schlezinger, R.D. White, J.J. Stegeman, Oxidative inactivation of cytochrome P4501A stimulated by 3,3%,4,4%-tetrachlorobiphenyl: production of reactive oxygen by vertebrate CYP1As, Mol. Pharmacol. 56 (1999) 588–597. [34] J.J. Schlezinger, W.D.J. Struntz, J.J. Stegeman, Oxidative inactivation of cytochrome P4501A and production of reactive oxygen stimulated by halogenated aryl hydrocarbon receptor agonists (submitted). [35] R.E. Farrell, RNA Methodologies. A Laboratory Guide for Isolation and Characterization, Academic Press, San Diego, 1993. [36] B. Bachmann, W. Luke, G. Hunsmann, Improvement of PCR amplified DNA sequencing with the aid of detergents, Nuc. Acids Res. 18 (1990) 1309. [37] D.G. Higgins, P.M. Sharp, CLUSTAL: a package for performing multiple sequence alignment on a microcomputer, Gene 73 (1988) 237–244. [38] J.J. Stegeman, R.L. Binder, A. Orren, Hepatic and extrahepatic microsomal electron transport components and mixed-function oxygenases in the marine fish Stenotomus 6ersicolor, Biochem. Pharmacol. 28 (1979) 3431–3439. [39] S.S. Park, H. Miller, A.V. Klotz, P.J. Kloepper-Sams, J.J. Stegeman, H.V. Gelboin, Monoclonal antibodies to liver microsomal cytochrome P-450E of the marine fish Stenotomus chrysops (scup): cross-reactivity with 3-methylcholanthrene induced rat cytochrome P-450, Arch. Biochem. Biophys. 249 (1986) 339–350. [40] F. Deryckere, F. Gannon, A 1-h minipreparation technique for extraction of DNA-binding proteins from animal tissues, Biotechniques 16 (1994) 405. [41] D.W. Ballard, W.H. Walker, S. Doerre, et al., The 6-rel oncogene encodes a kB enhancer binding protein that inhibits NF-kB function, Cell 63 (1990) 803–814. [42] D.W. Ballard, E. Bohnlein, J.A. Hoffman, et al., Activation of the interleukin-2 receptor alpha gene: regulatory role for DNA-protein interactions flanking the kB enhancer, New Biol. 1 (1989) 183–192. [43] M. Meyer, H.L. Pahl, P.A. Baeuerle, Regulation of the transcription factors NF-kB and AP-1 by redox changes, Chem. Biol. Interact. 91 (1994) 91–100. [44] P. Pfisterer, A. Annweiler, C. Ullmer, L.M. Corcoran, T. Wirth, Differential transactivation potential of Oct1 and Oct2 is determined by additional B-cell-specific activities, EMBO J. 13 (1994) 1655–1663. [45] S.F. Altschul, W. Gish, W. Miller, E.W. Myers, D.J. Lipman, Basic local alignment search tool, J. Mol. Biol. 215 (1990) 403–410. 156 J.J. Schlezinger et al. / Chemico-Biological Interactions 126 (2000) 137–157

[46] R.D. White, D. Shea, A.R. Solow, J.J. Stegeman, Induction and post-transcriptional suppression of hepatic cytochrome P4501A1 by 3,3%,4,4%-tetrachlorobiphenyl, Biochem. Pharmacol. 53 (1997) 1029–1040. [47] J.H. Leu, M.S. Lee, K.T. Chen, G.D. Change, C.K. Chou, C.J. Huang, Genomic structure of carp mitogen-activated protein kinase kinase 1 gene, Biochim. Biophys. Acta 1306 (1996) 133–136. [48] P.A. Baeuerle, T. Henkel, Function and activation of NF-kB in the immune system, Annu. Rev. Immunol. 12 (1994) 141–179. [49] C. Kunsch, S.M. Ruben, C.A. Rosen, Selection of optimal kappaB/Rel DNA-binding motifs: interaction of both subunits of NF-kappaB with DNA is required for transcriptional activation, Mol. Cell. Biol. 12 (1992) 4412–4421. [50] D.E. Cressman, L.E. Greenbaum, B.A. Haber, R. Taub, Rapid activation of post-hepatectomy factor/nuclear factor kappa B in hepatocytes, a primary response in the regenerating liver, J. Biol. Chem. 269 (1994) 30429–30435. [51] A.A. Beg, W.C. Sha, R.T. Bronson, S. Ghosh, D. Baltimore, Embryonic lethality and liver degeneration in mice lacking the RelA component of NF-kB, Nature 376 (1995) 167–170. [52] I. Rusyn, L. Cohn, R. Schoonhoven, J.A. Swenberg, R.G. Thurman, Corn oil rapidly activates nuclear factor kB (NF-kB) in Kupffer cells in vivo, Toxicol. Sci. 48 (1999) 237. [53] S.J. Stohs, Oxidative stress induced by 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD), Free Rad. Biol. Med. 9 (1990) 79–90. [54] C.E.W. Sulentic, M.P. Holsapple, N.E. Kaminski, A structure activity relationship of various polychlorinated dibenzo-p-dioxin congeners: inhibition of immunoglobulin expression versus AhR- mediated enzyme induction, Toxicol. Sci. 48 (1999) 179. [55] Y. Tian, S. Ke, M.S. Denison, A.B. Rabson, M.A. Gallo, Ah receptor and NF-kB interactions, a potential mechanism for dioxin toxicity, J. Biol. Chem. 274 (1999) 510–515. [56] X.P. Pang, N.S. Ross, M. Park, G.J. Juillard, T.M. Stanley, J.M. Hershman, Tumor necrosis factor-a activates nuclear factor kB and induces manganous superoxide dismutase and phosphodi- esterase mRNA in human papillary thyroid carcinoma cells, J. Biol. Chem. 267 (1992) 12826– 12830. [57] K.S. Yao, P.J. O’Dwyer, Involvement of NF-kB in the induction of NAD(P)H:quinone oxidoredu- tase (DT-diaphorase) by hypoxia, oltipraz and mitomycin C, Biochem. Pharmacol. 49 (1995) 275–282. [58] Q.W. Xie, Y. Kashiwabara, C. Nathan, Role of transcription factor NF-kB/Rel in induction of nitric oxide synthase, J. Biol. Chem 269 (1994) 4705–4708. [59] E.L. Kwak, D.A. Larochelle, C. Beaumont, S.V. Torti, F.M. Torti, Role for NF-kBinthe regulation of ferritin H by tumor necrosis factor-a, J. Biol. Chem. 270 (1995) 15285–15293. [60] G. Zhou, M.T. Kuo, NF-kB-mediated induction of mdr1b expression by insulin in rat hepatoma cells, J. Biol. Chem. 272 (1997) 15174–15183. [61] E.G. Schuetz, J.D. Schuetz, M.T. Thompson, R.A. Fisher, J.R. Madariage, S.C. Strom, Phenotypic variability in induction of P-glycoprotein mRNA by aromatic hydrocarbons in primary human hepatocytes, Mol. Carcinog. 12 (1995) 61–65. [62] G.C. Yeh, J. Lopaczynska, C.M. Poore, J.A. Phang, A new functional role for P-glycoprotein: efﬂux pump for benzo[a]pyrene in human breast cancer MCF-7 cells, Cancer Res. 52 (1992) 6692–6695. [63] R.J. Grumont, S. Gerondakis, Structure of a mammalian c-rel protein deduced from the nucleotide sequence of murine cDNA clones, Oncogene Res. 4 (1989) 1–8. [64] E. Brownwell, N. Mittereder, N.R. Rice, A human rel proto-oncogene cDNA containing an Alu fragment as a potential coding exon, Oncogene 4 (1989) 935–942. [65] S. Ghosh, A.M. Gifford, L.R. Riviere, P. Tempst, G.P. Nolan, D. Baltimore, Cloning of the p50 DNA binding subunit of NF-kB: homology to rel and dorsal, Cell 62 (1990) 1019–1029. [66] G.P. Nolan, S. Ghosh, H.C. Liou, P. Tempst, D. Baltimore, DNA binding and IkB inhibition of the cloned p65 subunit of NF-kB, a rel-related polypeptide, Cell 64 (1991) 961–969. [67] R.P. Ryseck, P. Bull, M. Takamiya, et al., a new Rel family transcription activator that can interact with p50-NF-kB, Mol. Cell. Biol. 12 (1992) 674–684. J.J. Schlezinger et al. / Chemico-Biological Interactions 126 (2000) 137–157 157

[68] M. Kieran, V. Blank, F. Logeat, et al., The DNA binding subunit of NF-kB is identical to factor kBF1 and homologous to the rel oncogene product, Cell 62 (1990) 1007–1018. [69] V. Bours, P.R. Burd, K. Brown, et al., A novel mitogen-inducible gene product related to p50/p105-NF-kB participates in transactivation through a kB site, Mol. Cell. Biol. 12 (1992) 685–695. [70] S.M. Ruben, P.J. Dillon, R. Schreck, et al., Isolation of a rel-related human cDNA that potentially encodes the 65-kD subunit of NF-kB, Science 251 (1991) 1490–1493. [71] S.M. Ruben, J.F. Klement, T.A. Coleman, M. Maher, C.H. Chen, C.A. Rosen, I-Rel: a novel rel-related protein that inhibits NF-kB transcriptional activity, Genes Dev. 6 (1992) 745–760.