Characterization of a Subset of the Basic-Helix-Loop-Helix-PASSuperfamily That Interacts with Components of the DioxinSignaling Pathway*

(Received for publication, April 1, 1996, and in revised form, December 3, 1996)

John B. Hogenesch‡§, William K. Chan‡, Victoria H. Jackiw‡, R. Clark Brown‡,Yi-Zhong Gu§, Marilyn Pray-Grant¶, Gary H. Perdew¶, and Christopher A. Bradfield§i

From the ‡Department of Molecular Pharmacology and Biological Chemistry, Northwestern University Medical School,Chicago, Illinois 60611, the §McArdle Laboratory for Cancer Research, University of Wisconsin Medical School,Madison, Wisconsin 53706, and the ¶Department of Veterinary Science, Pennsylvania State University,State College, Pennsylvania 16802

In an effort to better understand the mechanism oftoxicity of 2,3,7,8-tetrachlorodibenzo-p-dioxin, we em-ployed an iterative search of human expressed sequencetags to identify novel basic-helix-loop-helix-PAS (bHLH-PAS) proteins that interact with either the Ah receptor(AHR) or the Ah receptor nuclear translocator (ARNT).We characterized five new “members of the PAS super-family,” or MOPs 1–5, that are similar in size and struc-tural organization to the AHR and ARNT. MOPs 1–4have N-terminal bHLH and PAS domains and C-terminalvariable regions. MOP5 contained the characteristicPAS domain and a variable C terminus; it is possiblethat the cDNA contains a bHLH domain, but the entireopen reading frame has yet to be completed. Coimmu-noprecipitation studies, yeast two-hybrid analysis, andtransient transfection experiments demonstrated thatMOP1 and MOP2 dimerize with ARNT and that thesecomplexes are transcriptionally active at defined DNAenhancer sequences in vivo. MOP3 was found to associ-ate with the AHR in vitro but not in vivo. This observa-tion, coupled with the fact that MOP3 formed tighterassociations with the 90-kDa heat shock protein thanthe human AHR, suggests that MOP3 may be a condi-tionally active bHLH-PAS protein that requires activa-tion by an unknown ligand. The expression profiles ofthe AHR, MOP1, and MOP2 mRNAs, coupled with theobservation that they all share ARNT as a common di-meric partner, suggests that the cellular pathways me-diated by MOP1 and MOP2 may influence or respond tothe dioxin signaling pathway.

The AHR,1 ARNT, SIM, and PER are the founding members

of an emerging superfamily of regulatory proteins (1–4). TheAHR and ARNT are dimeric partners that transcriptionallyup-regulate genes involved in the metabolism of xenobioticcompounds. The AHR is activated by a number of widespreadenvironmental pollutants, such as the prototypical agonist,TCDD. In the absence of ligand, the AHR is primarily cytosolicand functionally repressed, presumably as the result of its tightassociation with HSP90 (5). Current models suggest that ago-nist binding initiates translocation of the receptor complex tothe nucleus and concomitantly weakens the AHRzHSP90 asso-ciation. Within the nucleus, HSP90 is displaced, and the AHRdimerizes with its partner ARNT, resulting in a bHLH-PASdimer with binding specificity for enhancer elements upstreamof gene products that metabolize foreign chemicals (6). In Dro-sophila, SIM is the master regulator of midline cell lineage inthe embryonic nervous system (7). Genetic, in vitro, and in vivostudies suggest that SIM may also dimerize with an ARNT-likeprotein in Drosophila and regulate enhancer sequences presentin the sim, slit, and Toll structural genes (7–10). The Drosoph-ila PER protein plays a role in the maintenance of circadianrhythms. PER has been shown to form heterotypic interactionswith a second Drosophila protein TIM in vivo, and homotypicinteractions with the ARNT molecule in vitro (11, 12).

The distinguishing characteristic of these proteins is a 200–300 stretch of amino acid sequence similarity known as thePAS domain. In the AHR, the PAS domain has been shown toencode sites for agonist binding, surfaces to support dimeriza-tion with other PAS domains, as well as surfaces that formtight interactions with HSP90 (1). In addition to the PASdomain, the AHR, ARNT, and SIM also harbor a bHLH motifthat plays a primary role in dimer formation. The bHLH motifis found in a variety of transcription factors that utilize homo-typic interactions to dimerize and regulate various aspects ofcell growth and differentiation (1, 13–15). Dimerization speci-ficity is conferred by sequences within both the bHLH anddeterminants within secondary interaction surfaces, such asthe “leucine zipper” or “PAS” domains (1, 12, 16). Interestingly,these dimerization surfaces also appear to restrict pairing towithin a given bHLH protein superfamily, thus minimizingcross-talk between important cellular pathways (17).

* This work was supported by the Burroughs Wellcome Foundation;by National Institutes of Health Grants P30-CA07175, ES05703,ES05660, ES05700, and GM08061; and by a postdoctoral fellowshipsponsored by the Colgate-Palmolive Company. The costs of publicationof this article were defrayed in part by the payment of page charges.This article must therefore be hereby marked “advertisement” in ac-cordance with 18 U.S.C. Section 1734 solely to indicate this fact.

The nucleotide sequence(s) reported in this paper has been submittedto the GenBankTM/EBI Data Bank with accession number(s) U29165and U51625–U51628.

i To whom all correspondence should be addressed: McArdle Labora-tory for Cancer Research, 1400 University Ave., Madison, WI 53706.Tel.: 608-262-2024; Fax: 608-262-2824; E-mail: bradfield@oncology.wisc.edu.

1 The abbreviations used are: AHR, Ah receptor; ARNT, Ah receptornuclear translocator; bHLH, basic-helix-loop-helix; PAS, PER/ARNT/SIM homology domain; TCDD, 2,3,7,8-tetrachlorodibenzo-p-dioxin;PER, the protein product of the Drosophila period gene; SIM, theprotein product of the Drosophila single-minded gene; HIF1a, hypoxia-

inducible factor 1 a (also referred to as MOP1 in this paper); HIF1b,hypoxia-inducible factor 1 b (also referred to as ARNT in this paper);EST, expressed sequence tag; BLAST, basic local alignment search tool;PCR, polymerase chain reaction; ORF, open reading frame; MOP, mem-bers of PAS superfamily; IPTG, isopropyl-b-D-thiogalactopyranoside;bNF, b-napthoflavone; bp, base pair(s); kb, kilobase pair(s); MOPS,4-morpholinepropanesulfonic acid; PAGE, polyacrylamide gelelectrophoresis.

THE JOURNAL OF BIOLOGICAL CHEMISTRY Vol. 272, No. 13, Issue of March 28, pp. 8581–8593, 1997© 1997 by The American Society for Biochemistry and Molecular Biology, Inc. Printed in U.S.A.

This paper is available on line at http://www-jbc.stanford.edu/jbc/ 8581

At the time we began this work, the AHR and ARNT werethe only mammalian bHLH-PAS proteins that had been iden-tified (see “Discussion”). Because other bHLH protein familiesutilize multiple homotypic interactions to provide fine controlin the regulation of various gene batteries, we predicted thatadditional bHLH-PAS proteins existed in the mammalian ge-nome and that a subset of these proteins would dimerize witheither the AHR or ARNT. We propose that identification ofsuch partners and determination of their pairing rules are thefirst steps in characterizing the potential points of cross-talkbetween different bHLH-PAS-mediated signaling pathways, aswell as understanding the pleiotropic responses initiated bypotent AHR agonists like TCDD. These observations led us toinitiate a search for additional family members from librariesof human ESTs.

EXPERIMENTAL PROCEDURES

Search Strategy—The bHLH-PAS domains of the huAHR, huARNT,drSIM, and the PAS domain of drPER were used as query sequences inBLASTN searches of the GenBank™ data base between December of1994 and October of 1995, using the following default values: database 5 NR (NR, non-redundant subset), expect 5 10, word length 5 12(18). Preliminary experiments comparing AHR and PER led us to definecandidate ESTs as those “hits” that yielded scores of 150 or higher. Asa method to confirm the similarity of these EST sequences to knownbHLH-PAS proteins, each candidate EST subsequently was comparedwith the NR subset of GenBank™ using the BLASTX program, ma-trix 5 blosum 62, word length 5 3. Only ESTs that retrieved knownbHLH-PAS proteins by this method of confirmation were furthercharacterized.

Oligonucleotide Sequences—Sequences of oligonucleotides are givenbelow. In cases where the oligonucleotide was used in gel shift assays,the 6-bp target sequence is underlined.

Cloning Strategy—In an effort to obtain extended open readingframes for each EST, an anchored-PCR strategy was employed to am-plify additional flanking sequence from a variety of commercial cDNAlibraries that were constructed in the phagemid Lambda Zap (tissues:HepG2, fetal brain, and skeletal muscle) (Stratagene, La Jolla, CA)(Table I) (19). The resulting PCR products were subjected to agarose gelelectrophoresis, transferred to a nylon membrane, and analyzed byhybridization with a 32P-labeled probe generated from the correspond-ing parent EST plasmid (Table I). After autoradiography, the positivePCR products were purified by gel electrophoresis and cloned using thepGEM-T vector system (Promega, Madison, WI). Dideoxy sequencingwas performed to characterize each positive clone (20).

Plasmid Construction for Expression in Vivo—Sequence informationfrom each EST was used to design PCR primers for the amplification ofcDNA from commercially available libraries. Expression plasmids wereconstructed by standard protocols (21). For a summary of clone desig-nations, PCR primers, DNA templates, and GenBank™ accession num-bers, refer to Table I. A brief description follows.

MOP1 Expression Vectors—Oligonucleotides OL404 and OL365 wereused as primers in a PCR to amplify a 970-bp fragment from a HepG2cell cDNA library. This fragment was cloned into the pGEM-T vector inthe T7 orientation and designated PL439. To generate pGMOP1, theSalI/XhoI fragment of hbc025 was subcloned into SalI-digested PL439.To increase transcription efficiency of the MOP1 cDNA, pGMOP1 wasdigested with KpnI and SacI and this fragment subcloned into thecorresponding sites of pSputk generating PL415 (Stratagene, La Jolla,CA) (22). The complete ORF of the MOP1 cDNA was amplified using thePCR and oligonucleotides OL425 and OL536. This fragment was di-gested with BamHI and ligated into the BamHI site in the pSportpolylinker (Life Technologies, Inc.). This plasmid was designatedPL611.

MOP2 Expression Vectors—The PCR was employed using OL477 andOL450 to amplify a 931-bp MOP2 fragment from a HepG2 cDNA li-brary. This fragment was cloned into pGEM-T in the SP6 orientationand designated PL424. Using OL560 and OL590, PCR amplificationfrom this same library yielded a 39 fragment of the MOP2 cDNA. Thisfragment was cloned into pGEM-T in the SP6 orientation and wasdesignated PL445. PL424 was digested with SalI and EcoRI and thefragment ligated into a SalI/EcoRI-digested PL445 to generate a fullORF MOP2 expression vector designated PL447. The complete ORF ofthe MOP2 cDNA was cloned into pSport as follows; PL447 was digestedwith SacII, treated with the Klenow fragment of DNA polymerase I in

the presence of dNTPs, and subsequently digested with SalI. Thisfragment was purified and ligated into pSport, digested with HindIII,repaired with Klenow, then digested with SalI. This construct wasdesignated PL477.

MOP3 Expression Vectors—Using the primers OL145 and OL489and a human fetal brain cDNA library as template, the PCR was usedto obtain a 1380-bp fragment. This fragment was isolated and clonedinto pGEM-T as above, and this plasmid designated PL487. A fragmentof MOP3 was obtained by the PCR using Pfu polymerase (Stratagene),primers OL657 and OL689, and PL487 as template. To obtain a full-length MOP3 cDNA fragment, the megaprimer fragment obtainedabove was used in the PCR against oligonucleotide OL611 using IM-AGE clone 50519 as a template (23). This product was cloned intopGEM-T in the SP6 orientation as above and designated PL425.

MOP4 Expression Vectors—Using primers OL520 and OL145 and aHepG2 cDNA library as template, the PCR was performed to isolate a59 fragment of the MOP4 cDNA. This fragment was cloned in the T7orientation of pGEM-T and designated PL448. The cDNA insert of thephage clone F9047 (from C. C. Liew, University of Toronto, Toronto,Canada) was amplified by the PCR using oligonucleotides OL418 andOL419 and subcloned into the pGEM-T vector (24). This clone wasdesignated PL420. An EcoRI fragment of PL448 was isolated andcloned into a partially EcoRI digested PL420. This clone was subjectedto the PCR using oligonucleotides OL698 and OL146, the fragmentcloned into the pGEM-T vector, and designated PL545.

MOP5 Expression Vectors—The PCR was used to obtain a 1260-bpfragment of the MOP5 gene using oligonucleotides OL685 and OL686and IMAGE clone 42596 as template. This fragment was purified andsubcloned into the pGEM-T vector as above in the SP6 orientation. Thisplasmid was designated PL528, and subsequently digested with SalIand partially digested with NcoI. This fragment was ligated into NcoI/SalI-cut pSputk, and the resulting vector designated PL554.

Hypoxia-responsive Luciferase Reporters—The epo-luciferase plas-mid, pGL2EPOEN, was constructed as follows. The hypoxia-responsiveenhancer from the 39 region of the EPO gene was amplified by PCRusing oligonucleotides OL499 and OL500 and human genomic DNA astemplate (amplified fragment corresponds to nucleotides 127–321 asreported in the EPO structural gene sequence found in GenBank™accession no. GBL16588). This fragment was digested with KpnI andNheI and cloned into the corresponding sites of the plasmid pGL2-Promoter (Promega).

Antibody Production—Antisera against MOP1, MOP2, AHR, andARNT were prepared in rabbits using immunization protocols thathave been described previously (25, 26). Crude antisera were chosenfor use in all coimmunoprecipitation experiments, and the preim-mune sera from the same rabbit served to preclear the samples. ForMOP1, the plasmid hbc025 was digested with EcoRI and the 604-bpfragment was treated with the Klenow fragment of DNA polymerase1 in the presence of dNTPs and cloned into the SmaI site of thehistidine tag fusion vector pQE-32 (Qiagen, Chatsworth, CA). Thisclone, designated PL377, was transformed by electroporation intoM15(REP4) cells for expression under IPTG induction. The expressedprotein was purified from 8 M urea using nickel-nitrilotriacetic acid-agarose, extensively dialyzed against 25 mM MOPS, pH 7.4, 100 mM

KCl, and 10% glycerol before its use as an immunogen. Antiserumproduced against this protein was designated R3752. For AHR, thehuman cDNA clone PL71 (27) was digested with BamHI and clonedinto the corresponding site of the histidine fusion vector pQE31(Qiagen). The AHR protein fragment was expressed and purifiedexactly as described for MOP1 (above). Antiserum produced againstthis protein was designated R2891. For MOP2, a SacI/PstI fragmentof PL445 was cloned into SacI/PstI cut pQE-31 to generate PL456.This clone, designated PL456, was transformed into M15(REP4) cellsand the protein expressed under IPTG induction. The histidine-tagged fusion protein was first extracted in guanidine hydrochloride,dialyzed extensively, and purified on nickel-nitrilotriacetic acid-aga-rose as above. Antiserum produced against this protein was desig-nated R4064. ARNT-specific antiserum was raised against huARNTprotein purified from baculovirus as described previously (28).

Northern Protocol—Multiple tissue Northern blots containing 2 mg ofpoly(A)1 mRNA prepared from human heart, brain, placenta, lung,liver, skeletal muscle, kidney, and pancreas were probed with randomprimed cDNA fragments using an aqueous hybridization protocol (Clon-tech, Palo Alto, CA). Hybridization solution contained 5 3 SSPE (0.75 M

NaCl, 50 mM NaH2PO4, 5 mM Na2EDTA, pH 7.4), 2 3 Denhardt’ssolution (0.04% w/v Ficoll 400, 0.04% w/v polyvinylpyrrolidone, 0.04%w/v bovine serum albumin), 0.5% SDS, and 100 mg/ml heat denaturedsalmon sperm DNA. The blot was prehybridized for 3–6 h at 65 °C, the

bHLH-PAS Subset That Interacts with AHR or ARNT8582

hybridization solution was changed, and 1–5 3 106 cpm/ml of a randomprimed cDNA fragment was added. Samples were hybridized overnightat 65 °C, washed twice with 2 3 SSC (0.3 M NaCl, 30 mM Na3 citrate, pH7.0), 0.5% SDS at room temperature, once with 1 3 SSC, 0.1% SDS atthe hybridization temperature, and once with 0.1 3 SSC, 0.1% SDS atthe hybridization temperature.

Yeast Two-hybrid Analysis—A modified yeast interaction trap wasemployed to identify those MOPs that could interact with the AHR orARNT in vivo. LexAMOP chimeras were constructed to fuse the bHLH-PAS domains of the MOP proteins with the DNA binding domain of thebacterial protein LexA amino acids 1–202) (29). To amplify the regioncorresponding to the bHLH-PAS domains of MOP1, OL715 and OL716were employed in the PCR using PL415 as template. To amplify the

region corresponding to the bHLH-PAS domains of MOP2, OL717 andOL718 were employed in the PCR using PL447 as template. To amplifythe region corresponding to the bHLH-PAS domains of MOP3, OL719and OL720 were employed in the PCR using PL486 as template. Toamplify the region corresponding to the bHLH-PAS domains of MOP4,OL721 and OL722 were employed in the PCR using PL545. Since amore detailed domain map existed for the AHR, a construct was madewith a fine deletion of the transactivation domain. The N-terminalportion of the AHR was amplified by the PCR using oligonucleotidesOL180 and OL124 and pmuAHR as template (30). This product wasdigested with KpnI and SalI, and cloned into the corresponding sites ofpSG424 (31). This clone was designated PL187. The 39 end of the AHRcDNA was amplified by PCR using oligonucleotides OL201 and OL202

OL21 59-CGAGGTCGACGGTATCG-39OL22 59-TCTAGAACTAGTGGATC-39OL124 59-CCCAAGCTTACGCGTGGTCTTTGAAGTCAACCTCACC-39OL145 59-AGCTCGAAATTAACCCTCACTAAAGG-39OL176 59-CGGGATCCTTACACATTGGTGTTGGTACAGATGATGTACTC-39OL180 59-GCGTCGACTGATGAGCAGCGGCGCCAACATCACC-39OL201 59-GATAAGAATGCGGCCGCAGATCTGGGTCCGAAGCACACG-39OL202 59-CATTACTTATCTAGAGCTCG-39OL226 59-CGGGATCCTCATGGCGGCGACTACTGCCAACC-39OL365 59-GACAGTTGCTTGAGTTTCAACC-39OL386 59-TTATGAGCTTGCTCATCAGTTGCC-39OL387 59-CCTCACACGCAAATAGCTGATGG-39OL392 59-CCGCTCGAGTGATGAGCAGCGGCGCCAACATCACC-39OL393 59-CCGCTCGAGTGGCAGCTACAGGAATCCACC-39OL404 59-GCGGTACCGGGACCGATTCACCATGGAG-39OL414 59-TCGAGCTGGGCAGGGTACGTGGCAAGGC-39OL415 59-TCGAGCCTTGCCACGTACCCTGCCCAGC-39OL418 59-GTAAAACGACGGCCAGT-39OL419 59-GGAAACAGCTATGACCATG-39OL425 59-GGGATCCACCGATTCACCATGGAG-39OL443 59-TCGAGCTGGGCAGGGTGCGTGGCAAGGC-39OL444 59-TCGAGCCTTGCCACGCACCCTGCCCAGC-39OL445 59-TCGAGCTGGGCAGGTCACGTGGCAAGGC-39OL446 59-TCGAGCCTTGCCACGTGACCTGCCCAGC-39OL447 59-TCGAGCTGGGCAGGTTGCGTGGCAAGGC-39OL448 59-TCGAGCCTTGCCACGCAACCTGCCCAGC-39OL450 59-TACTGGCCACTTACTACCTGACC-39OL456 59-AACCAGAGCCATTTTTGAGACT-39OL477 59-GCTCTAGAGGCCACAGCGACAATGACAGC-39OL479 59-GATCGGAGGTGTTCTATGAGC-39OL489 59-TTAGGATGCAGGTAGTCAAACA-39OL496 59-GTTCTCCATGGACCAGACTGA-39OL499 59-CGGGTACCCTGGGCCCTACGTGCTGTCTC-39OL500 59-CGGCTAGCCTCTGGCCTCCCTCTCCTTGATGA-39OL514 59-CTGGGAGCCTGCCTGCCTTCA-39OL520 59-CCCAAGGAGAGGCGTGAT-39OL536 59-GCGGATCCGTCAGTTAACTTGATCCAAAGC-39OL540 59-GGGATCCTCGTCGCCACTG-39OL541 59-ATGCAGTACCCAGACGGATTTC-39OL560 59-TGCACGGTCACCAACAGAG-39OL561 59-TTGCCAGTCGCATGATGGA-39OL565 59-CTGAACAGCCATCCTTAG-39OL568 59-AGCTTGCCCTACGTGCTGTCTCAG-39OL569 59-AATTCTGAGACAGCACGTAGGGCA-39OL590 59-AGAGGTGCTGCCCAGGTAGAA-39OL611 59-CAATGATGAGGGAAACACTG-39OL657 59-CGGGATCCCGTCAACTGGAGATGAGCAAGGAG-39OL665 59-CTGCAAAAATCCGATGACCTCTT-39OL681 59-CGGGCAGCAGCGTCTTC-39OL682 59-GCGTCCGCAGCCCCAGTTG-39OL683 59-TTCAATGTTCTCATCAAAGAGC-39OL684 59-GAACAGTTTTATAGATGAATTGGC-39OL685 59-GGAATTCCATGGGTCTCTCACAGGTGGAGATGAC-39OL686 59-GGAATTCCTCAGTGATGGTGATGGTGATGGTCCCCCTTCCGCGGCAGCCTGG-39OL689 59-GAGGTGTTTCAATTCATCGTCT-39OL715 59-GGGATCCGTGACCGATTCACCATGGAG-39OL716 59-CTGCAGGTCACACAACGTAATTCACACA-39OL717 59-GGGATCCGTATGACAGCTGACAAGGAG-39OL718 59-GGTCGACGTCACAGGACGTAGTTGACACA-39OL719 59-GAATCCATGAGCAAGGAGGCCGTG-39OL720 59-GGTCGACGTCAAACAACAGTGTTAGTTGA-39OL721 59-GGGATGCGTATGGATGAAGATGAGAAAGAC-39OL722 59-GGTCGACGCTAGACCGAGTGTGTGCA-39

SEQUENCES 1–64

bHLH-PAS Subset That Interacts with AHR or ARNT 8583

and pmuAHR as template. This product was digested with NotI andcloned into the corresponding site of pSGAhND166 (30). This clone wasdesignated PL188. PL188 was digested with KpnI and XbaI, and thisfragment cloned into the corresponding sites of PL187. This clone wasdesignated PL204. A cDNA fragment of the AHR was amplified by thePCR using OL392 and OL393 and PL204 as template. This product wascloned using the pGEM-T system and designated pGTAHRDTAD. Thisconstruct was digested with XhoI and this fragment ligated into SalI-cut pBTM116 (32). This construct was designated pBTMAHR. Lex-AARNT was constructed by PCR using oligonucleotides OL226 andOL176 and PL87 as template. The PCR product was cloned intopGEM-T as above, and the BamHI fragment cloned into a BamHI-digested pGBT9 vector (Clontech). This construct was cut with BamHIand subcloned into a BamHI-digested pBTM116. This construct wasdesignated LexAARNT. Following amplification these products werepurified and cloned into the pGEM-T vector. These clones were desig-nated PL537 (MOP1), PL538 (MOP2), PL539 (MO3), and PL540(MOP4). These plasmids were digested with BamHI/PstI (PL537),BamHI/SalI (PL538 and PL540), and EcoRI/SalI (PL539), and thesefragments ligated into the appropriately digested pBTM116. Theseclones were designated LexAMOP1, LexAMOP2, LexAMOP3, and Lex-AMOP4, respectively. Full-length expression plasmids harboring theAHR and ARNT were constructed as follows: PL104 (pSporthuAHR)was cut with SmaI, the insert purified and subcloned into SmaI site ofpCW10, and this plasmid was designated PL317. This clone was di-gested with SmaI and subcloned into a SmaI-cut pRS305, and this clonedesignated pRSAHR. The ARNT cDNA (PL101) was digested with NotIand XhoI and cloned into the corresponding sites of pSGBMX1. Thisplasmid was designated PL371, and subsequently was digested withNotI and XhoI, and this fragment cloned into the corresponding sites ofpSGBCU11. This clone was designated PL574. The LexAMOP fusionprotein constructs were cotransformed with a yeast expression vectorcontaining the full-length AHR or ARNT into L40, a yeast strain con-taining integrated lacZ and HIS3 reporter genes. As controls, Lex-AAHR and LexAARNT were cotransformed with AHR and ARNT. Thestrength of interaction was visually characterized by 5-bromo-4-chloro-3-indolyl b-D-galactoside plate assays, performed after 3 days of growthon selective medium (33). To provide quantitation of the interactionstrength, multiple colonies from yeast harboring each bHLH-PAS com-bination colonies were grown overnight in liquid medium. Liquid cul-tures were grown for 5 h and assayed for lacZ activity using theGalacto-Light chemiluminescence reporter system (Tropix, Bedford,MA). To determine the effect of AHR agonists on these interactions,yeast were also grown on plates or in liquid culture with and without 1mM bNF (34).

To ensure expression of each bHLH-PAS construct, Western blotanalysis was performed using antibodies raised against the LexA DNAbinding domain. Yeast extracts were prepared from 15-ml overnightcultures derived from multiple colonies of yeast expressing each Lex-AMOP fusion protein. Cultures were subjected to centrifugation at1200 3 g for 5 min, and the pellet was resuspended in 500 ml of 6 M

guanidinium HCl, 0.1 M sodium phosphate buffer, 0.01 M Tris, pH 8.0.This suspension was transferred to a fresh Eppendorf tube containing500 ml of acid washed glass beads (Sigma), and mixed on the maxsetting in a Bead-Beater (BioSpec, Bartlesville, OK) for 3 min at 4 °C.

The samples were cleared by centrifugation at 14,000 3 g, and 400 ml ofsupernatant was precipitated with 400 ml of 10% trichloroacetic acid onice. After clearing by centrifugation at 14,000 3 g for 20 min at 4 °C, theextracts were resuspended in SDS loading buffer and subjected toSDS-PAGE analysis. Following electrophoresis, proteins were trans-ferred to nitrocellulose membrane and detected with LexA antiserumand secondary antibodies linked to alkaline phosphatase by standardprotocols (35).

Transient Transfection of Hep3b Cells—pGL2EPOEN was cotrans-fected with pSport, PL611 (pSportMOP1), or PL477 (pSportMOP2)using the Lipofectin protocol (Life Technologies, Inc.). Briefly, the ex-pression vector was mixed with the Epo reporter and the b-galactosid-ase control plasmid pCH110 (Clontech) and incubated with cells in asix-well plate for 5 h at 37 °C in serum-free medium. Following incu-bation, fresh medium was added, and the cells were incubated with orwithout 75 mM cobalt chloride at 37 °C overnight before the cells wereharvested. Cell extraction and luciferase assays were performed accord-ing to manufacturer’s protocols (Promega), and b-galactosidase assayswere performed using the Galacto-Light assay according to manufac-turer’s protocols (Tropix).

Coimmunoprecipitation with HSP90—Each MOP construct was invitro translated in the presence of [35S]methionine in a TNT-coupledtranscription/translation system (Promega). HSP90 immunoprecipita-tion assays were performed with monoclonal antibody 3G3p90 or acontrol IgM antibody, TEPC 183 (Sigma) essentially as described (36).Each immunoprecipitation was subjected to SDS-PAGE, and the result-ing gel was dried. The level of radioactivity in each coprecipitatedprotein band was quantified on a Bio-Rad GS-363 Molecular ImagerSystem. The amount of bHLH-PAS protein immunoprecipitated withthe HSP90 antibody is presented as a percentage of the amount ofmurine AHR immunoprecipitated in parallel assays.

RESULTS

EST Search

Our initial BLAST searches in December 1994 were per-formed with the bHLH-PAS or PAS domains of all familymembers known at that time (AHR, ARNT, SIM, and PER). Inthese searches we identified an EST clone, hbc025, derivedfrom human pancreatic islets (Table I) (37). To confirm thissimilarity, we performed a BLASTX search, comparing hbc025to the GenBank™ data base and found that this sequence wasmost homologous to Drosophila SIM (Fig. 1). This EST clonewas designated MOP1. The bHLH-PAS domains of all familymembers, including MOP1, were again searched from May toOctober of 1995. Human ESTs that recorded BLASTN scoresabove 150 were again retrieved and confirmed using theBLASTX algorithm. This routine resulted in the discovery ofsix ESTs with significant homology to the bHLH-PAS domainsof known members (Fig. 1, Table I).

TABLE IMOP cDNA clone information

Row 1, clones (in parentheses) containing the candidate ESTs were requested from their laboratory of origin. Row 2, the GenBank accessionnumber for each original EST is indicated. Row 3, oligonucleotides used in library screening. Sequence information generated from this originalclone was used to design oligonucleotides for use in an anchored-PCR strategy, whereby gene-specific and vector-specific primers were used toamplify 59 and 39 portions of the cDNA. Vector specific primers corresponded to modified T3 (59, OL145) or T7 (39, OL146) primers. A matrix ofgene-specific primers against annealing temperature (50–65 °C) was attempted for each clone, generally leading to at least one successful reaction.Row 4, the cDNA libraries from which additional sequence of positive clones was identified. Row 5, size of ORFs. We define a complete ORF by thepresence of an in-frame stop codon 59 to a methionine codon that lies within a Kozak consensus sequence for translational initiation. The 39 endof open reading frames are defined by the presence of an in-frame termination codon. An asterisk (*) denotes a clone which does not meet thesecriteria (see text). Row 6, GenBank™ accession numbers of the final MOP cDNAs are given.

MOP1/HIF1a MOP2 MOP3 MOP4 MOP5

Laboratory of origin (clone designation) Bell (hbc025) IMAGE (67043) IMAGE (23820, 50519) Liew (F9047, PL420) IMAGE (42596)EST GenBank™ accession no. T10821 T70415 T77200, H17840 R58054 R67292Gene-specific oligonucleotides used in

PCROL365 (59) OL456 (59)

OL496 (39)OL514 (39)OL541 (39)

OL489 (59) OL520 (59) OL540 (59)

Library screened HepG2 HepG2 Fetal brain HeLa HepG2ORF size (amino acids) 826 870 624 642* 412*Final cDNA GenBank™ accession no. U29165 U51626 U51627 U51625 U51628

bHLH-PAS Subset That Interacts with AHR or ARNT8584

cDNA Cloning

To more completely characterize the similarities and domainstructures of the candidate clones, an anchored-PCR strategywas employed to obtain additional flanking cDNA sequenceusing phagemid libraries as a template. Comparison of aminoacid sequences of these bHLH-PAS proteins is displayed in Fig.2. Upon characterization of the open reading frames, it waslearned that two of these ESTs (F06906 and T77200) corre-sponded to the same gene product (Table I, Fig. 1). Thus, wedesignated these remaining five unique cDNAs as “members ofPAS superfamily” or MOPs 1–5. The PCR strategy providedwhat appeared to be the complete ORFs of MOP1, MOP2, andMOP3 based upon the following criteria. 1) At their 59 endsthese clones contain an initiation methionine codon (AUG)downstream of an in-frame stop codon, and 2) at their 39 endsthese clones contain an in-frame stop codon followed by noobvious open reading frames. In addition, the nucleotide se-quences flanking of the MOP1 and MOP2 most 59 AUG codons(see GenBank™ accession nos. U29165 and U51626) are inreasonable agreement with the proposed optimal context fortranslational initiation, i.e. CCACCAUGG (38, 39). Using thesame anchored-PCR technique, we were unable to obtain thecomplete open reading frames of MOP4 or MOP5 (19). We did

identify a potential start methionine for MOP4 and the 39 stopcodon for MOP5 (Fig. 2). Our preliminary designation of theMOP4 start methionine is tentative and is based on its prox-imity to the start methionines of MOP1, MOP2, MOP3, AHR,and SIM (Fig. 2) (1, 3).

Tissue-specific Expression

To characterize the tissue-specific expression patterns of theMOP mRNAs, Northern blots of poly(A)1 RNA from eight hu-man tissues were probed with random primed cDNA restrictionfragments (Fig. 3). Single transcripts of 3.6 kb (MOP1), 6.6 kb(MOP2), and 3.2 kb (MOP3) were detected. Expression levels ofeach mRNA varied significantly between tissues, with MOP1being highest in kidney and heart, MOP2 highly expressed inplacenta, lung, and heart, and MOP3 highly expressed in skel-etal muscle and brain. No detectable message was detected forMOP4 or MOP5 by our Northern blot protocol.

Identification of Novel AHR or ARNT Partners

Interaction of MOPs with the AHR and ARNT in Vitro: Co-immunoprecipitation experiments—We first performed coim-munoprecipitation experiments to determine if MOPs 1–4 hadthe capacity to interact with either the AHR or ARNT in vitro.

FIG. 1. Schematic representation of a generic bHLH-PAS member and the corresponding region where EST “hits” occurred. Top,schematic of a generic bHLH-PAS family member. The hatched box represents the bHLH region, the overlined area represents the PAS domainwith the characteristic “A” and “B” repeats in white. The variable C terminus is boxed in white. A bold line representing the region in a genericbHLH-PAS member where the homology occurs is indicated next to the original GenBank™ accession number for each identified EST (MOP1 5T10821, MOP2 5 T70415, MOP3 5 T77200 and F06906, MOP4 5 R58054, MOP5 5 R67292; see Table I). Bottom, results of BLASTX searches areshown with the most homologous PAS members. The numbers adjacent to the query sequence (top) are relative nucleic acid positions of thehomology. The numbers adjacent to the subject sequence (bottom) are relative amino acid positions of the most homologous bHLH-PAS protein.

bHLH-PAS Subset That Interacts with AHR or ARNT 8585

These proteins were expressed in a reticulocyte lysate systemin the presence of [35S]methionine and then incubated in thepresence or absence of the AHR or ARNT. Complex formationwas assayed by coimmunoprecipitation with AHR- or ARNT-specific antisera, followed by quantitation of coimmunoprecipi-

tated 35S-labeled MOP by phosphoimage analysis (Fig. 4). In-teractions were identified by a reproducible increase in anAHR- or ARNT-dependent precipitation of MOP protein. Be-cause we have observed considerable variability in this coim-munoprecipitation assay, each experiment was performed atleast three times. A representative result is presented in Fig. 4.

In the AHR interaction studies (Fig. 4, top), we observed thatMOP3 was coimmunoprecipitated with AHR. The positive con-trol, ARNTzAHR interaction, was also reproducible, but weaker(Fig. 4, top). Neither MOP1, MOP2 or MOP4 could be shown tointeract with the AHR by this protocol. The ARNT proteindisplayed a broad range of interactions and was shown tocoimmunoprecipitate with AHR (positive control), MOP1 andMOP2, but not MOP3 or MOP4 (Fig. 4, bottom).

Interaction of MOPs with the AHR and ARNT in Vivo: YeastTwo-hybrid Experiments—To determine if MOPzAHR orMOPzARNT complexes could form in vivo, a modified interac-tion trap was employed (40, 41) (Fig. 5). Fusion proteins wereconstructed in which the DNA binding domain of the bacterialrepressor, LexA, was fused to the bHLH-PAS domains of theMOP proteins (Fig. 5, panel A). The bHLH-PAS domains werechosen because they harbor both the primary and secondarydimerization surfaces of this family of proteins and they typi-cally do not harbor transcriptional activity that would interferewith this assay (35). Interactions were tested by cotransforma-tion of each LexAMOP construct with either the full-lengthAHR or ARNT into the L40 yeast strain, which harbors anintegrated lacZ reporter gene driven by multiple LexA operator

FIG. 2. Amino acid sequence and multiple alignment of the PAS domains. The amino acid sequence including a CLUSTAL alignment ofthe bHLH-PAS domains is depicted. The CLUSTAL alignment was performed using the MEGALIGN program (DNASTAR, Madison, WI) with aPAM250 weight table using the following parameters: Ktuple 5 1, Gap Penalty 5 3, Window 5 5. Amino acid boundaries for the residuesencompassing the bHLH and PAS domains of the MOPs were defined based on previous observations. The bHLH domain is boxed, while the basicregion is specified by a vertical line. The PAS domain is underlined, while the “A” and “B” repeats of the PAS domain are boxed. Consensus (60%)residues in the PAS domain are denoted with an asterisk.

FIG. 3. Northern blot analyses of MOPs. Each lane consists of 2 mgof twice selected poly(A)1 mRNA prepared from human heart, brain,placenta, lung, liver, skeletal muscle, kidney, and pancreas (Clontech).The blots from top to bottom were probed for: MOP1, MOP2, MOP3, andGAPDH; exposure times were 5 days, 1 day, 3 days, and 8 h, respec-tively. Random primed cDNA fragments were as follows: MOP1, a604-bp EcoRI fragment of hbc025; MOP2, a 649-bp EcoRI/HindIII frag-ment of IMAGE clone 67043; MOP3, a 1121-bp HindIII fragment ofIMAGE clone 50519; MOP4, a 696-bp SacI fragment of clone PL420 (seeTable I); MOP5, a 539-bp BamHI/HindIII fragment of IMAGE clone42596; and GAPDH (rat), a 1.3-kb PstI fragment of PL89 (pGAPDH).IMAGE clone designations are given in Table I.

bHLH-PAS Subset That Interacts with AHR or ARNT8586

sites (32). In this system, LexAMOP fusions that interact withAHR or ARNT drive expression of the lacZ reporter gene.

We assessed the relative strength of these interactions byboth a direct lacZ plate assay and by quantitation of the re-porter activity in a liquid culture (Fig. 5B). In all cases, thesetwo methods of detection were equivalent (data not shown). Totest the validity of this model system as a method to detectbHLH-PAS interactions, LexAAHR and LexAARNT constructswere cotransformed with either the full-length ARNT or AHR.In these control experiments, we were able to demonstrate thespecificity of AHRzARNT interaction and its dependence on thepresence of the agonist bNF. The LexAAHRzARNT interactionin the presence of bNF was 913-fold above background, whilethe LexAARNTzAHR interaction in the presence of bNF was14-fold above background. Both combinations showed ligandinducibility. The LexAAHRzARNT interaction in the presenceof bNF was 6.4-fold greater than LexAAHRzARNT in the ab-sence of ligand, while the LexAARNTzAHR interaction in thepresence of bNF was 2.0-fold over LexAARNTzAHR in theabsence of ligand. Despite our ability to readily detect theagonist-induced LexAARNTzAHR interaction in the two-hybridsystem, we were unable to detect any LexAMOP that could

FIG. 4. In vitro coimmunoprecipitation of radiolabeled MOPs1–4 with AHR (top) or ARNT (bottom). 35S-Labeled in vitro trans-lated ARNT, AHR, or MOPs 1–4 (2–10 ml of lysate reaction) wereincubated with 10 ml of in vitro translated, unlabeled AHR or 5 ml of invitro translated, unlabeled ARNT (plus lanes) or an equivalent amountof uncharged reticulocyte lysate (minus lanes) for 10 min at 30 °C.Preimmune serum (10 ml) was then added to preclear the samples. After15 min at room temperature, 1 ml of immunoprecipitation buffer (50mM HEPES, pH 7.4, containing 150 mM KCl, 10% glycerol, 1 mM EDTA,10 mM b-mercaptoethanol, and 0.1% Triton X-100) and 30 ml of proteinA-trisacryl resin (Pierce) were added. After gentle mixing for 15 min at4 °C, the samples were subjected to centrifugation and the superna-tants were transferred to new vials containing 10 ml of AHR (R2891) orARNT-specific (R3611) antibodies. After 30 min at room temperature,30 ml of protein A-trisacryl resin was again added and the samples weregently mixed for 60 min at 4 °C. After centrifugation, the pellets werewashed three times with immunoprecipitation buffer and then ana-lyzed by SDS-PAGE. The gels were dried and analyzed byphosphoimaging.

FIG. 5. Yeast two-hybrid analysis. In vivo interaction of MOPswith dioxin signaling pathway. Panel A, schematic of AHR, ARNT, andLexA fusion constructs. Panel shows a schematic of the AHR, with thePAS domain (black) with the characteristic “A” and “B” repeats (white),the bHLH domain (striped), and the variable C terminus (white). Thetranscriptionally active glutamine rich domain is indicated with “Q”(shaded box). LexA fusion proteins are indicated with the N terminus ofLexA DNA-binding protein fused to bHLH-PAS domains of MOPs 1–4,and ARNT. The LexAAHR construct contains the bHLH-PAS domainsand the C terminus minus the transcriptionally active Q-rich region(see “Experimental Procedures”). Panel B, relative interaction of LexAfusion proteins with the AHR or ARNT. Galacto-Light assays wereperformed on yeast extracts prepared from colonies expressing LexAM-OPs, LexAAHR, or LexAARNT and ARNT or AHR in the presence andabsence of 1 mM bNF. Assays were performed in triplicate, and then therelative light units normalized to the LexAAHRzARNT 1bNF conditioninternally (set as 100%). The stippled bars represent LexA fusion pro-teins co-expressed with the full-length AHR in the presence of 1 mM

bNF, the striped bars represent the fusion proteins co-expressed withthe full-length AHR in the absence of ligand, the shaded bars indicatethe fusion proteins co-expressed with the full-length ARNT, and theopen bar indicates LexAAHR co-expressed with full-length ARNT in thepresence of 1 mM bNF. Panel C, Western blot analysis of LexA fusionproteins. Yeast extracts were prepared from colonies transformed withthe LexA fusion proteins. These extracts were subjected to SDS-PAGE;LexAMOP1 (lane 1), LexAMOP2 (lane 2), LexAMOP3 (lane 3), Lex-AMOP4 (lane 4), LexAAHR (lane 5), and LexAARNT (lane 6).

bHLH-PAS Subset That Interacts with AHR or ARNT 8587

interact with the AHR (Fig. 5B). That is, none of the LexAMOPfusion proteins appeared to interact with cotransformed AHRand drive lacZ expression in the absence or presence of ligand(Fig. 5B).

Two of four MOP proteins tested were found to interact withARNT in the two-hybrid assay. Both the LexAMOP1 and Lex-AMOP2 interactions with full-length ARNT were extremelyrobust, 36- and 28-fold above background, respectively (Fig.5B). When compared with the LexAAHRzARNT interaction inthe presence of bNF, the LexAMOP1-ARNT and LexAMOP2-ARNT interactions were 24% and 69% as intense. These differ-ences in LexAMOP1zARNT and LexAMOP2zARNT interactioncould be attributed to differences in expression levels or tosubtle differences in chimera construction. To control for rela-tive expression of the LexAMOP fusions, protein extracts wereprepared and Western blot analysis was performed with LexA-specific antiserum (Fig. 5C). We observed expression for eachLexAMOP fusion proteins, indicating that negative resultswith LexAMOP3 and LexAMOP4 are not due to lack ofexpression.

DNA Binding and Specificity—Prompted by the observationthat MOP1 and ARNT and MOP2 and ARNT specifically in-teract, we next examined the ability of MOPzARNT dimericcomplexes to bind those DNA response elements recognized byother bHLH-PAS protein complexes. Reports from a number oflaboratories have demonstrated that bHLH-PAS dimers canbind to a variety of DNA elements: “DRE,” TNGCGTG (42);“CME,” ACGTG (8); “SAE,” GTGCGTG (10); and “E-box,”CANNTG (9, 10). Using a gel-shift assay, we observed thatMOP1zARNT complexes specifically bound CACGTG andTACGTG (Fig. 6A, lanes 6 and 7), while the complex failed tobind GTGCGTG, TTGCGTG, and a non-palindromic E-box,CATGTG (lanes 8–10). Previous reports have demonstratedthat ARNT homodimers are capable of binding the CACGTGsequence in vitro and that this complex can drive reporter geneexpression in vivo (9, 10). Our results suggest that theMOP1zARNT dimeric complex binds the CACGTG oligonucleo-tide with a higher affinity than either MOP1 or ARNT alone(Fig. 6A, lanes 11–14). MOP1 failed to form a productive DNAbinding complex with the AHR with any of the bHLH-PASfamily response elements (Fig. 6A, lanes 2–5). As a comparisonof MOP1zARNT and MOP2zARNT DNA binding, we provideresults from gel shift assays using a double-stranded oligonu-cleotide containing a core TACGTG hexad binding site (Fig.6B). We observed that both MOP1zARNT and MOP2zARNTbound the TACGTG-containing oligonucleotide with approxi-mately equal capacity and that neither ARNT, nor MOP1, norMOP2 could bind this DNA sequence alone. As additional con-trols, we confirmed the presence of the MOP proteins in thecomplex by showing that antisera raised against these proteinsretarded the mobility of the complex (Fig. 6B, lanes 6–11).

Interaction of MOPs with HSP90—In an effort to assess theability of MOPs to interact with HSP90, we performed coim-munoprecipitation assays with HSP90-specific antibodies.Given the remarkable stability of the HSP90 complex with the

FIG. 6. Gel-shift analysis of the MOP1-MOP2 and ARNT inter-action in vitro. A, 32P-labeled oligonucleotides corresponding to theknown bHLH-PAS response elements were used in each analysis. 32P-oligos indicates the core sequences of the given radiolabeled, double-stranded oligonucleotide used in that reaction. All mixing experimentswere performed with equimolar amounts of each protein unless other-wise indicated. The proteins were coincubated at 30 °C for 30 min,followed by the addition of poly(dI-dC) (100 ng, 10 min at room temper-ature) and 5 3 104 cpm of the 32P-labeled and annealed oligonucleo-tides. All oligonucleotides share identical sequences flanking these vari-able core sequences. Arrows indicate the position of the MOP1zARNTand AHRCD516zARNT. Lane 1, the migration of the AHRCD516zARNTz32P-TTGCGTG complexes as a positive control. Lanes 2–5, gelshifts performed with various radiolabeled oligonucleotides andequimolar mixtures of MOP1 and AHRCD516. Lanes 11–14, gel-shiftedband observed in the presence of MOP1zARNT, MOP1 alone, ARNTalone, and 2-fold excess of ARNT with the 32P-TCACGTG radiolabeledoligonucleotide. B, MOP1 and MOP2 were expressed and quantitated inreticulocyte lysates as described under “Experimental Procedures.”ARNT was expressed and purified from a baculovirus expression sys-tem as described under “Experimental Procedures.” OligonucleotidesOL568 and OL569 containing a TACGTG core sequence were labeledand annealed as described under “Experimental Procedures.” EachDNA binding reaction contained 0.6 fmol of MOP1 or MOP2 and 20 fmolof ARNT. The proteins were coincubated at 30 °C for 30 min, followedby the addition of poly(dI-dC) (100 ng, 10 min at room temperature) and5 3 104 cpm of the 32P-labeled and annealed oligonucleotides. Lanes

1–5, binding reactions contained MOP1 and ARNT (lane 1), MOP2and ARNT (lane 2), MOP1 alone (lane 3), MOP2 alone (lane 4) or ARNTalone (lane 5). Lanes 6–11, to confirm the role of MOP1 and MOP2 inARNTzDNA complex formation, supershift experiments with antiserumraised against MOP1 (HIF1-Ab) or MOP2 (MOP2-Ab) were performed(lanes 6, 8, 9, and 11). Preimmune serum (PI) from the same rabbits wasused in lanes 7 and 10 as a control for nonspecific inhibition of thebinding by serum. Antisera were added after the oligonucleotide incu-bation step and an additional incubation of 10 min at room temperaturewas added. Samples were loaded onto a 0.3 3 Tris borate-EDTA buffernondenaturing polyacrylamide gel (5%), dried, and subjected toautoradiography.

bHLH-PAS Subset That Interacts with AHR or ARNT8588

AHR from the C57BL/6J mouse, we used this receptor speciesas a reference and compared all interactions relative to it. Asadditional controls, we immunoprecipitated ARNT and the hu-man AHR as negative and positive controls, respectively. De-spite our ability to readily detect huAHRzHSP90 interactions,we were unable to detect ARNT, MOP2, or MOP5 interactionswith HSP90 (Fig. 7). In contrast, huAHR, MOP1, MOP3, andMOP4 all immunoprecipitated with HSP90-specific antisera.MOP3 formed the tightest interaction with HSP90, followed bythe huAHR, MOP4, and MOP1 (71%, 53%, 31%, and 17%,respectively.

DISCUSSION

Hypothesis—Our hypothesis was that additional bHLH-PASproteins are encoded in the mammalian genome and that someof these proteins are involved in mediating the pleiotropicresponse to potent AHR agonists like TCDD. This idea arosefrom the observation that other bHLH superfamilies employmultiple dimeric partnerships to control complex biological pro-cesses, such as myogenesis (MyoD/myogenin), cellular prolifer-ation (Myc, Max, Mad), and neurogenesis (achaete-scute/daughterless) (43–45). The observation that bHLH proteinsoften restrict their dimerization to within members of the samegene family (i.e. “homotypic interactions”) and that this restric-tion may occur as the result of constraints imposed by bothprimary (e.g. bHLH) and secondary dimerization surfaces (e.g.leucine zippers and PAS), prompted us to screen for additionalbHLH-PAS proteins and test each protein for its capacity tointeract with either the AHR or ARNT. The ultimate objectiveof our search was to identify MOPs that were physiologicallyrelevant partners of either the AHR or ARNT in vivo. Ourprediction was that such proteins might respond to or modulatethe AHR signaling pathway and thus provide mechanistic in-sights into TCDD toxicity.

Expressed Sequence Tag Approach—In an effort to rapidlyidentify expressed genes, Venter and colleagues (46) developedthe EST approach, whereby a cDNA library is constructed andrandomly selected clones are sequenced from both vector arms.These partial sequences, generally 200–400 bp, are depositedin a number of computer data bases that can be readily ana-lyzed using a variety of search algorithms. In the past year, theIMAGE (Integrated Molecular Analysis of Genomes and theirExpression) Consortium has deposited over 300,000 humanESTs, generated from different tissues and developmental timeperiods into publicly accessible data bases, identifying approx-

imately 40,000 unique cDNA clones2 (47). The availability ofthese sequences and plasmids harboring their correspondingcDNA clones provided the impetus to identify novel members ofthe bHLH-PAS family by nucleotide homology screening ofavailable EST data bases.

When we began these studies, the human AHR and ARNTand the Drosophila SIM and PER were the only PAS proteinsthat had been described. Therefore, we used the nucleotidesequences encoding their PAS domains as query sequences inBLASTN searches of the available EST data bases. Using thisstrategy in an iterative fashion and confirming each hit with aBLASTX search (nucleotide against protein search), we identi-fied five MOP cDNAs (Fig. 1). Using PCR, we were able toobtain the complete ORF of MOPs 1–3, and extensive butincomplete ORFs of MOP4 and 5. The inability to obtain thecomplete ORFs of MOP4 and MOP5 appears to be related to thelow copy number of their mRNAs in the tissues examined.Interestingly, MOPs 1–4 displayed adjacent bHLH-PAS do-mains and variable C termini. This domain structure is iden-tical to that observed in the AHR, ARNT, and SIM. MOP5contained a consensus PAS domain, but our inability to amplifyits 59 end did not allow us to determine if it is also a bHLHprotein.

While this work was in progress, Wang and colleagues iden-tified two factors involved in cellular response to hypoxia,HIF1a and HIF1b. These proteins are identical to MOP1 andARNT, respectively (48). Thus, of the five MOPs we havecloned, four have not been previously characterized. For con-sistency in this report, we use the term MOP1 for HIF1a.Because of historical precedent and since the term MOP1 wasmeant to be temporary nomenclature, we will refer to this cloneas HIF-1a in future reports.

Since cDNAs encoding the complete open reading frames forMOPs 1–3 were available, our initial studies focused on theseproteins. Although these data should be considered prelimi-nary, MOP4 and MOP5 were included in some studies. Sinceour MOP4 clone contained the sequences analogous to thoseinvolved in the dimerization, transcriptional activation, andDNA binding of other bHLH-PAS proteins, this cDNA wasincluded in most interaction analyses (30, 35). Since our MOP5clone did not harbor a bHLH domain (required for dimericinteractions with other MOPs), but did encode a complete PASdomain (analogous to the AHR region required for HSP90interaction), this clone was included in the HSP90 coimmuno-precipitation experiment.

Tissue-specific Expression—To provide a preliminary indica-tion of the biological relevance of each MOP cDNA to the AHRsignaling pathway, we performed Northern blot analyses todetermine which MOPs displayed overlapping tissue-specificexpression profiles with either the AHR or ARNT mRNAs. Weobserved that each MOP mRNA displayed a unique tissue-specific distribution, with MOP1 being highest in kidney andheart, MOP2 highly expressed in placenta, lung, and heart, andMOP3 highly expressed in skeletal muscle and brain. Previousstudies conducted in our laboratory indicated that the humanARNT is ubiquitously expressed with highest levels in skeletalmuscle and placenta, while the human AHR is most prevalentin placenta, lung, and heart and lower levels in brain, liver, andskeletal muscle (27, 49). The observation that these bHLH-PASproteins are coexpressed in a variety of tissues supports theidea that cross-talk between these signaling pathways may beoccurring in vivo and that multiple tissue-specific interactionsmay be taking place. In particular, the observation that AHR

2 This estimate was provided by the IMAGE Consortium (http://www-bio.llnl.gov/bbrp/I.M.A.G.E./iachievements.html and ftp://humpty.llnl.gov/pub/EST/IndexReport).

FIG. 7. Interaction of MOPs with HSP90. In vitro expressed ra-diolabeled muAHR, huAHR, and MOPs1–5 were immunoprecipitatedwith HSP90-specific antisera. Following precipitation, SDS-PAGEanalysis was performed, the gel dried, and exposed to a phosphorimaging plate for analysis. The values presented are the average of twoseparate experiments, and compared with the muAHR (set at 100%).Values from these two assays differed by less than 5%.

bHLH-PAS Subset That Interacts with AHR or ARNT 8589

and MOP2 have coincident expression profiles in human tis-sues suggested to us that these proteins should be among thefirst candidates to be screened for signaling pathway interac-tion. An additional and equally important interpretation ofthese unique MOP expression profiles is that unidentified part-ners exist for these bHLH-PAS proteins and that they regulatea number of undescribed biological pathways.

Interaction Strategy—Our interaction screening strategywas based on functional data and the detailed domain mapsavailable for the AHR and ARNT. An important assumptionused in the design and interpretation of our studies is thatsome of the MOPs may be constitutively active in vivo (likeARNT) and others may be conditionally active, possibly requir-ing ligand-activation to dimerize in vivo (like AHR). We choseto employ coimmunoprecipitation as an initial interactionscreen for a number of reasons. First, AHR and ARNT-specificantibodies are available that have been shown to precipitateAHRzARNT complexes. This suggests that if MOPzAHR orMOPzARNT interactions occurred in vitro, that these sameantibodies would recognize and precipitate such complexes.Second, data from a number of laboratories, using independ-ently derived antibodies, indicate that coimmunoprecipitationof AHRzARNT complexes is independent of AHR-ligand (50,51). This observation suggests that AHR or ARNT interactionswith conditional MOP proteins might still be identified bycoimmunoprecipitation even in the absence of knowledge abouthow to activate a conditional MOP (e.g. identification of itsligand).

As a secondary screen to characterize interacting MOPs, weemployed a yeast interaction trap commonly referred to as the“two-hybrid assay” (40). Support for use of this system comesfrom our previous observation that LexAAHR chimeras arefunctional in yeast and provide a good model of AHR signalingand ARNT interaction (34). In addition, this method providesan independent confirmation of those interactions identified bycoimmunoprecipitation and also provides a demonstration thatinteractions can occur in vivo. One advantage of comparingtwo-hybrid results with coimmunoprecipitation results is thatit may reveal conditional MOPs that require activation prior todimerization. An example of this can be seen with the AHR andARNT. In the absence of ligand, the AHR appears to resideprimarily in the cytosol and ARNT appears to be primarilynuclear (26, 35). This compartmentalization appears to be partof a cellular mechanism to prevent inappropriate interaction ofthese proteins and minimize constitutive activity of the com-plex. As a LexA chimera in yeast, the AHR is repressed untiladdition of ligand (34). Therefore, in vivo systems such as thetwo-hybrid assay may yield negative results for conditionalMOP proteins in the absence of the factors required for theiractivation (e.g. ligands).

In light of the above considerations, our interpretation of thecoimmunoprecipitation and two-hybrid interaction results areas follows. First, since the MOP1zARNT and the MOP2zARNTinteractions were confirmed in two independent systems, theseinteractions appear to involve MOPs that have constitutiveactivity. Second, the observation that MOP3 interacts with theAHR in vitro, but not in vivo, suggests that MOP3 may be aconditional MOP that has the capacity to interact with theAHR or other MOPs in vivo upon its activation by a ligand (thisidea gained support from HSP90 interaction studies below).The suspicion that MOP3 is a conditional bHLH-PAS protein,coupled with the observation that MOP3 and AHR have dis-parate expression profiles, led us to delay study of this inter-action until we learn how to activate MOP3 or until we haveevidence that these two proteins are expressed in the same celltype. Finally, our observation that ARNT can form dimers with

two out of four MOPs examined suggests that ARNT is apromiscuous bHLH-PAS partner that may be a focus of cross-talk between different MOP signaling pathways. The multiplic-ity of ARNT partnerships is supported by recent observationsfrom a number of laboratories (9, 10, 48).

MOP1 and MOP2 Interactions with ARNT—The concord-ance of the coimmunoprecipitation and two-hybrid data led usto pursue the MOP1zARNT and MOP2zARNT interactions fur-ther. Given the pairing rules deduced from the interactionstudies described above, we next attempted to determine if theMOP1zARNT and MOP2zARNT complexes bound specific DNAsequences in vitro. Earlier reports indicated that the basicregion of each bHLH partner generates specificity for a distinctDNA half-site of at least 3 bp (10, 16). Data from a number oflaboratories have indicated that the ARNT protein displaysspecificity for the 39-GTG half-site of the hexad target se-quence, 59-NNCGTG-39, where 59-NNC is the half-site of theARNT partner (10, 52). To determine the half-site specificity ofthe MOP1 protein when complexed with ARNT, we used gelshift analysis with oligonucleotides representing the responseelements of all known bHLH-PAS proteins. These preliminaryexperiments indicated that MOP1zARNT complex had greatestaffinity for the 59-CACGTG and 59-TACGTG sites (Fig. 6A).

At the time these DNA binding specificity studies were beingcompleted, a report appeared that defined the biological role ofMOP1. Studies by Semenza and colleagues demonstrated thata protein complex, termed HIF1, regulates hypoxia-responsivegenes such as EPO and VEGF, and is composed of a dimer ofHIF1a and HIF1b subunits (48). HIF1a and HIF1b RNAs wereboth shown to be up-regulated in response to hypoxia or certainagents like cobalt chloride or desferrioxamine that stimulatean upstream “heme sensor” (48). Moreover, these studies dem-onstrated that the HIF1 complex bound to TACGTG-containingenhancer elements that regulated the expression of hypoxia-responsive genes. These results were important for a numberof reasons. First, these results provided independent supportfor our screening approach by demonstrating that theMOP1zARNT complex we described did have biological rele-vance. Second, they confirmed our independently derived DNAbinding site for the MOP1zARNT complex. Third, they provideda logical approach to design a physiologically relevant reporterconstruct in our attempts to compare the interactions of MOP1and MOP2 with ARNT in vivo (see below).

Because the MOP1 and MOP2 basic regions differed by onlyone amino acid residue and since this residue is not thought tobe in a DNA contact position (53, 54), we hypothesized thatMOP2 would bind the same DNA half-site sequences as MOP1.To confirm this, we performed MOP2zARNT gel shift assaysusing a double-stranded oligonucleotide containing a coreTACGTG hexad binding site (Fig. 6B). We observed that bothMOP1zARNT and MOP2zARNT bound the TACGTG-containingoligonucleotide and that neither MOP1 nor MOP2 could bindthis sequence in the absence of ARNT. As additional controls,we confirmed the presence of the MOP1 and MOP2 proteins inthe complex by showing that antisera raised against theseproteins retarded the mobility of the complex (Fig. 6B, lanes6–11).

To assay MOP1zARNT and MOP2zARNT interactions in vivo,we constructed a luciferase reporter driven by the hypoxia-responsive TACGTG-containing enhancer from the humanEPO gene (55). Our transient expression experiments inHep3B cells consisted of cotransfection of this reporter withvector control, MOP1, or MOP2 in the presence or absence ofcobalt chloride to stimulate the hypoxia heme sensor (Fig. 8)(56). ARNT has been shown previously to be expressed inHEP3B cells (55). This experiment confirmed that the

bHLH-PAS Subset That Interacts with AHR or ARNT8590

TACGTG-containing enhancer sequence is responsive to cobaltand cotransfected MOP1 or MOP2 under normal oxygen ten-sion (Fig. 8). The transfected MOP1 construct appeared to beresponsive to hypoxia (3.5-fold over control), while the MOP2construct was only slightly responsive (1.2-fold) (Fig. 8). MOP2was more potent than MOP1 in driving expression of thisreporter gene both in the presence and absence of cobalt chlo-ride. This difference in efficacy of the MOP1 and MOP2 con-structs in driving expression of the reporter plasmid in Hep3Bcells could be explained by three possibilities. 1) The relativepotency of the MOP2 transactivation domain may be muchgreater than MOP1, 2) the relative expression of MOP2 my begreater in this transient expression system than MOP1, or 3)the MOP1 may be partially repressed in vivo by HSP90 whileMOP2 is not (see HSP90 discussion below). Given that ourMOP2 antisera are not useful in Western blots, we could not

assess the relative expression of the MOP1 and MOP2 clones inthis system. Thus, we cannot rule out more efficient expressionof MOP2 relative to MOP1.

MOP3 Is a Conditionally Active bHLH-PAS Protein—Datafrom a number of laboratories suggest that HSP90 is involvedin AHR signaling in vivo (34, 57). In vitro experiments suggestthat HSP90 is required for high affinity ligand binding andthat HSP90 “caps” the DNA binding domain, preventing theunliganded receptor from constitutively interacting with theARNT protein and binding DNA (58, 59). Moreover, a minimalregion shown to repress transactivation of constitutive AHRdeletion chimeras has been shown to mediate HSP90 binding(60). These observations suggest that HSP90 represses AHRactivity by inhibiting constitutive dimerization and by anchor-ing the receptor in the cytosol away from its nuclear dimericpartner ARNT. Upon ligand binding, the AHRzHSP90 complextranslocates to the nucleus where HSP90 dissociates from thecomplex and the AHR dimerizes with ARNT and binds DNA(6). Two lines of evidence suggest that MOP3, like the AHR,may be a conditionally active bHLH-PAS protein and that inthe absence of an unidentified cognate ligand, might be re-pressed and unable to dimerize in vivo. First, MOP3 interactswith HSP90 even more efficiently than the human AHR, sug-gesting that MOP3 may be functionally repressed or anchoredin the cytosol like the AHR (Fig. 7). Second, MOP3 interactswith AHR in the coimmunoprecipitation assay, but not in theyeast interaction trap (Fig. 4, top, and Fig. 5B). Similarly, theAHR interacts with ARNT in the coimmunoprecipitation assay,but interacts weakly, if at all, in the absence of ligand activa-tion (Fig. 4, bottom) (34).

Alternative explanations for the different MOP3zAHR inter-action results obtained from our in vivo or in vitro systemsmust also be considered. For example, the structure of MOP3may be different than the AHR and ARNT, such that position-ing of the LexA domain adjacent to the bHLH-PAS domain maysterically hinder dimerization surfaces within this protein orlead to improper subcellular localization or instability of thechimera. One example of the potential negative impact of con-text sensitivity in the two-hybrid system can be observed inFig. 5. The LexAAHRzARNT interaction is 14.7 times more

FIG. 8. Transcriptional activity of MOP1zARNT andMOP2zARNT complexes in vivo. Hep3B cells were cotransfected withpSport, pSportMOP1, or pSportMOP2 with the epo-luciferase reporterplasmid, pGL2EPOEN (see “Experimental Procedures” for details).Cells were then incubated at 37 °C overnight in the presence (blackbars) or absence (striped bars) of 75 mM cobalt chloride. The luciferaseunits are normalized to b-galactosidase units and represent mean 6S.E. of triplicate samples. The -fold difference is expressed as thecobalt-induced value over control for each transfected vector.

FIG. 9. Schematic comparison of homology of PAS family members. A dendrogram was prepared from the primary amino acid CLUSTALalignment above using the MEGALIGN program. The CLUSTAL alignment was performed using the MEGALIGN program (DNASTAR, Madison,WI) with a PAM250 weight table using the following parameters: Ktuple 5 1, Gap Penalty 5 3, Window 5 5. Amino acid boundaries for theresidues encompassing the bHLH and PAS domains of the MOPs were defined based on previous observations. The amino acid boundaries are asfollows: huMOP1/HIF1 (91–342), huMOP2 (90–342), huMOP3 (148–439), huMOP4 (87–350), huMOP5 (32–296), huAHR (117–385), huARNT(167–464), drSIM (82–356), drPER (232–496), bsKINA (27–248), huSRC-1 (115–365), muARNT2 (141–437), muSIM1 (83–331), muSIM2 (83–332),drSIMILAR (174–419), and drTRH (145–471). The scale at the bottom indicates number of amino acid residue substitutions. PAS family membersthat interact with HSP90, interact with the AHR, and interact with ARNT by the coimmunoprecipitation method are indicated by a 1, whereasmembers that do not interact are indicated by a 2 (see “Discussion”). An a denotes a bHLH-PAS member whose cDNA is not complete. Note thatthese interactions occur in vitro and may or may not be physiologically relevant. Where appropriate, the reference is included in parentheses.

bHLH-PAS Subset That Interacts with AHR or ARNT 8591

robust than the LexAARNTzAHR interaction. In addition, theLexAAHRzARNT interaction is more responsive to the AHRligand bNF than the LexAARNTzAHR combination (6.4-foldand 2.0-fold, respectively). This difference cannot be explainedby the relative transactivation potencies of the transactivationdomains of AHR and ARNT in yeast, and therefore must be aresult of context sensitivity. A final consideration is that coim-munoprecipitations may be capable of detecting weak interac-tions that cannot be maintained at the low cellular concentra-tions of the various MOPs. Thus, the MOP3zAHR dimerizationmay be too weak to occur in vivo. In this regard, we have previ-ously reported ARNTzARNT homodimers that bind specific DNAenhancer sequences in vitro, but they are weakly active, if at all,in vivo (10).

It is also important to note that MOP1 and MOP4 alsointeract with HSP90 in the coimmunoprecipitation assay, al-beit less strongly than MOP3 or human AHR (Fig. 7). Therelatively weak interaction of MOP1 with HSP90 may be anindication that this protein is partially repressed in vivo andthat it may have both constitutive and conditional activity.Such a phenomenon might explain why MOP1 has less tran-scriptional activity in our in vivo systems than MOP2, whichdoes not interact with HSP90 (Figs. 7 and 8). Finally, MOP4did not interact with the AHR or ARNT in either the coimmu-noprecipitation assay or the interaction trap. Although ourexperience with AHR indicates that interactions with condi-tional bHLH-PAS proteins can be observed by coimmunopre-cipitation assays, MOP4’s interaction with HSP90 may alsoindicate a requirement for activation and may inhibit the sen-sitivity of detecting interactions in vivo.

The bHLH-PAS Superfamily—Our experimental approachhas significantly expanded the number of known members ofthis emerging superfamily of transcriptional regulators andhas identified additional potential targets for TCDD signaling.During the preparation of this manuscript, five additionalmammalian bHLH-PAS proteins have been identified, HIF1a,SIM1, SIM2, ARNT2, and SRC-1 (48, 51, 61–64). To compareamino acid sequences of these proteins, we performed aCLUSTAL alignment with the bHLH-PAS domains of all theknown family members using a PAM250 residue weight table(65). The two most related members were MOP1/HIF1a andMOP2, which shared 66% identity in the PAS domain. A com-parison of these two proteins reveals only a single amino aciddifference in the basic region and 83% identity in the HLHregion. This sequence similarity is in agreement with our con-tention that MOP1/HIF1a and MOP2 function analogously,interacting with the same dimeric partners and binding similarenhancer sequences in vivo. A comparison of MOP3 and ARNTand a comparison of MOP5 and SIM reveal 40% and 38%identity in the PAS domain, respectively. The basic regions ofMOP3 and ARNT have only three substitutions, while the HLHdomains share 66% identity, again suggesting that the twoproteins may regulate similar or identical enhancer sequences(half-sites).

A CLUSTAL alignment of the C termini of these MOP pro-teins and the previously identified PAS members demonstratedthat these regions are not well conserved (data not shown) (1).This lack of conservation may indicate that the C termini ofthese genes have divergent functions, or that the functionsharbored in the C termini can be accomplished by a variety ofdifferent sequences. For example, the C termini of the AHR,ARNT, HIF1a, and SIM all harbor potent transactivation do-mains, yet display little sequence homology (35, 66, 67).

In an effort to characterize the evolutionary and functionalrelationships of these proteins, we performed a parsimonyanalysis to identify related subsets. A dendrogram represent-

ing the primary amino acid relationship between the PASdomains of these proteins is illustrated in Fig. 9. This figuresuggests that four major groups exist for eukaryotic PAS familymembers. The AHR, drSIMILAR, MOP1/HIF1a, MOP2,drTRACHEALESS, MOP5, and SIM exist in one group; ARNT,muARNT2, MOP3, and MOP4 exist in another; and PER andhuSRC-1 exist in their own groups. Interestingly, this patternreflects some of what is known functionally about the existingPAS members. Members from the AHR group and ARNT, butnot the PER group, have been shown to interact with HSP90(Fig. 7) (68). Only members from the ARNT group have beenshown to interact with the AHR (Fig. 4, top) (51). Finally,members from all groups have been shown to interact withARNT (Fig. 4, bottom) (10, 12, 48). The observation that ARNThas been shown to be capable of forming DNA binding ho-modimers as well as heterodimers with a number of previouslyidentified members of the bHLH-PAS family (at least in vitro),suggests that it plays a role in a number of biological processes(9, 10). Based on their similarity with ARNT, MOP3 and MOP4may be candidates for binding DNA as homodimers, or forinteracting with multiple bHLH-PAS members, possibly fromthe AHR group. PER is the only well characterized eukaryoticPAS protein so far identified that lacks a bHLH domain.huSRC-1 is the only PAS protein so far identified to interactwith members of the steroid receptor family. Future investiga-tion will focus on confirming the functional relationships ofthese MOPs, their pairing rules and DNA binding specificities.

Conclusion—In an effort to understand the molecular conse-quences to TCDD exposure, we have developed a protocol toidentify novel PAS proteins as well as to define their pairingrules. These studies bring the total number of the mammalianPAS family to 11. Using a number of assays for protein-proteininteractions and DNA binding specificity (10, 40), we have beenable to determine that both MOP1 and MOP2 are productivedimerization partners of the ARNT protein and that thesedimers recognize a TACGTG-containing enhancer element invivo. The validity of our system is supported by the character-ization of MOP1/HIF1a by an independent group that reachedthe same conclusions regarding dimerization and DNA bindingspecificity (48). These observations demonstrate the power ofthe approach and support its use in characterizing the emerg-ing superfamily of bHLH-PAS proteins that will be revealed byEST technology in the coming years.

In addition to the relevance of these data to TCDD signaling,it also appears to be revealing additional factors important tocellular responses to hypoxic stress. Our analysis indicatedthat HIF1a/MOP1 and MOP2 share a common dimeric partner,ARNT, and are capable of regulating a common battery ofgenes. This idea is supported by three lines of evidence; 1) bothMOP1 and MOP2 interact with ARNT as defined by coimmu-noprecipitation or two-hybrid assay, 2) they have similar DNAhalf-site specificities when complexed with ARNT, and 3) theyare both transcriptionally active from TACGTG enhancers invivo. The observation that HIF1a/MOP1 and MOP2 havemarkedly different tissue distributions suggests that these twoproteins may be regulating similar batteries of genes in re-sponse to different environmental or developmental stimuli.Alternatively, these proteins may be involved in restrictingexpression of certain groups of genes regulated by TACGTG-dependent enhancers. Finally, it is possible that MOP2 is asubunit of a “HIF1-like” complex (i.e. a “HIF2a”) that regulateshypoxia-responsive genes in a distinct set of tissues.

Acknowledgments—We thank the following for their generous con-tribution of reagents: for pBTM116, Paul Bartel and Stan Fields(SUNY, Stony Brook, NY); for L40, Rolf Sternglanz (SUNY, StonyBrook, NY); for hbc025, G. Belle (University of Chicago, Chicago, IL)(37); for F9047, C. C. Liew (University of Toronto, Toronto, Canada)

bHLH-PAS Subset That Interacts with AHR or ARNT8592

(24); for pCW10, P. Weil (Vanderbilt University School of Medicine,Nashville, TN) (69); for pSG424, M. Ptashne (Harvard University, Cam-bridge, MA) (31); for pRS305, P. Hieter (Johns Hopkins UniversitySchool of Medicine, Baltimore, MD) (70); for LexA antibodies, J. W.Little (University of Arizona, Tucson, AZ); and for pSGBMX1 andpSGBCU11, Stephen Goff (Ciba-Geigy Cooperation, Research TrianglePark, NC).

REFERENCES

1. Burbach, K. M., Poland, A., and Bradfield, C. A. (1992) Proc. Natl. Acad. Sci.U. S. A. 89, 8185–8189

2. Hoffman, E. C., Reyes, H., Chu, F. F., Sander, F., Conley, L. H., Brooks, B. A.,and Hankinson, O. (1991) Science 252, 954–958

3. Nambu, J. R., Lewis, J. O., Wharton, K. A., Jr., and Crews, S. T. (1991) Cell 67,1157–1167

4. Jackson, F. R., Bargiello, T. A., Yun, S. H., and Young, M. W. (1986) Nature320, 185–188

5. Perdew, G. H. (1992) Biochem. Biophys. Res. Commun. 182, 55–626. Swanson, H. I., and Bradfield, C. A. (1993) Pharmacogenetics 3, 213–2307. Muralidhar, M. G., Callahan, C. A., and Thomas, J. B. (1993) Mechanisms of

Development 41, 129–1388. Wharton, K., Jr., Franks, R. G., Kasai, Y., and Crews, S. T. (1994) Development

120, 3563–35699. Sogawa, K., Nakano, R., Kobayashi, A., Kikuchi, Y., Ohe, N., Matsushita, N.,

and Fujii-Kuriyama, Y. (1995) Proc. Natl. Acad. Sci. U. S. A. 92, 1936–194010. Swanson, H. I., Chan, W. K., and Bradfield, C. A. (1995) J. Biol. Chem. 270,

26292–2630211. Zeng, H., Qian, Z., Myers, M. P., and Rosbash, M. (1996) Nature 380, 129–13512. Huang, Z. J., Edery, I., and Rosbash, M. (1993) Nature 364, 259–26213. Neuhold, L. A., and Wold, B. (1993) Cell 74, 1033–104214. Amati, B., and Land, H. (1994) Curr. Opin. Genet. Dev. 4, 102–10815. Cabrera, C. V., and Alonso, M. C. (1991) EMBO J. 10, 2965–297316. Kadesch, T. (1993) Cell Growth & Diff. 4, 49–5517. Baxevanis, A. D., and Vinson, C. R. (1993) Curr. Opin. Genet. Dev. 3, 278–28518. Altschul, S. F., Gish, W., Miller, W., Myers, E. W., and Lipman, D. J. (1990) J.

Mol. Biol. 215, 403–41019. Innis, M. A., Gelfand, D. H., Sninsky, J. J., and White, T. J. (eds) (1990) PCR

Protocols: A Guide to Methods and Applications, Academic Press, SanDiego, CA

20. Sanger, F., Nicklen, S., and Coulson, A. R. (1977) Proc. Natl. Acad. Sci. U. S. A.74, 5463–5467

21. Sambrook, J., Fritsch, E. F., and Maniatis, T. (1989) Molecular Cloning: ALaboratory Manual, 2nd Ed., Cold Spring Harbor Laboratory Press, ColdSpring Harbor, NY

22. Falcone, D., and Andrews, D. W. (1991) Mol. Cell. Biol. 11, 2656–266423. Sarkar, G., and Sommer, S. S. (1990) Biotechniques 8, 404–40724. Hwang, D. M., Hwang, W. S., and Liew, C. C. (1994) J. Mol. Cell. Cardiol. 26,

1329–133325. Poland, A., Glover, E., and Bradfield, C. A. (1991) Mol. Pharmacol. 39, 20–2626. Pollenz, R. S., Sattler, C. A., and Poland, A. (1994) Mol. Pharmacol. 45,

428–43827. Dolwick, K. M., Schmidt, J. V., Carver, L. A., Swanson, H. I., and Bradfield, C.

A. (1993) Mol. Pharmacol. 44, 911–91728. Chan, W. K., Chu, R., Jain, S., Reddy, J. K., and Bradfield, C. A. (1994) J. Biol.

Chem. 269, 26464–2647129. Bartel, P., Chien, C. T., Sternglanz, R., and Fields, S. (1993) BioTechniques 14,

920–92430. Dolwick, K. M., Swanson, H. I., and Bradfield, C. A. (1993) Proc. Natl. Acad.

Sci. U. S. A. 90, 8566–857031. Sadowski, I., and Ptashne, M. (1989) Nucleic Acids Res. 17, 753932. Vojtek, A. B., Hollenberg, S. M., and Cooper, J. A. (1993) Cell 74, 205–21433. Bohen, S. P., and Yamamoto, K. R. (1993) Proc. Natl. Acad. Sci. U. S. A. 90,

11424–1142834. Carver, L. A., Jackiw, V., and Bradfield, C. A. (1994) J. Biol. Chem. 269,

30109–3011235. Jain, S., Dolwick, K. M., Schmidt, J. V., and Bradfield, C. A. (1994) J. Biol.

Chem. 269, 31518–31524

36. Fukunaga, B. N., Probst, M. R., Reisz-Porszasz, S., and Hankinson, O. (1995)J. Biol. Chem. 270, 29270–29278

37. Takeda, J., Yano, H., Eng, S., Zeng, Y., and Bell, G. I. (1993) Hum. Mol. Genet.2, 1793–1798

38. Kozak, M. (1986) Cell 44, 283–29239. Kozak, M. (1987) Nucleic Acids Res. 15, 8125–813240. Fields, S., and Song, O. (1989) Nature 340, 245–24641. Chien, C., Bartel, P. L., Sternglanz, R., and Fields, S. (1991) Proc. Natl. Acad.

Sci. U. S. A. 88, 9578–958242. Denison, M. S., Fisher, J. M., and Whitlock, J. J. (1989) J. Biol. Chem. 264,

16478–1648243. Rudnicki, M. A., and Jaenisch, R. (1995) Bioessays 17, 203–20944. Hurlin, P. J., Ayer, D. E., Grandori, C., and Eisenman, R. N. (1994) Cold

Spring Harbor Symp. Quant. Biol. 59, 109–11645. Skeath, J. B., and Carroll, S. B. (1994) FASEB J. 8, 714–72146. Adams, M. D., Kelley, J. M., Gocayne, J. D., Dubnick, M., Polymeropoulos, M.

H., Xiao, H., Merril, C. R., Wu, A., Olde, B., Moreno, R. F., Kerlavage, A. R.,McCombie, W. R., and Venter, J. C. (1991) Science 252, 1651–1656

47. Lennon, G., Auffray, C., Polymeropoulos, M., and Soares, M. B. (1996)Genomics 33, 151–152

48. Wang, G. L., Jiang, B. H., Rue, E. A., and Semenza, G. L. (1995) Proc. Natl.Acad. Sci. U. S. A. 92, 5510–5514

49. Carver, L. A., Hogenesch, J. B., and Bradfield, C. A. (1994) Nucleic Acids Res.22, 3038–3044

50. McGuire, J., Whitelaw, M. L., Pongratz, I., Gustafsson, J. A., and Poellinger, L.(1994) Mol. Cell. Biol. 14, 2438–2446

51. Hirose, K., Morita, M., Ema, M., Mimura, J., Hamada, H., Fujii, H., Saijo, Y.,Gotoh, O., Sogawa, K., and Fujii-Kuriyama, Y. (1996) Mol. Cell. Biol. 16,1706–1713

52. Reisz-Porszasz, S., Probst, M. R., Fukunaga, B. N., and Hankinson, O. (1994)Mol. Cell. Biol. 14, 6075–6086

53. Ellenberger, T., Fass, D., Arnaud, M., and Harrison, S. (1994) Genes & Dev. 8,970–980

54. Dang, C. V., Dolde, C., Gillison, M. L., and Kato, G. J. (1992) Proc. Natl. Acad.Sci. U. S. A. 89, 599–602

55. Semenza, G. L. (1994) Hematol. Oncol. Clin. N. Am. 8, 863–88456. Wang, G. L., and Semenza, G. L. (1993) Proc. Natl. Acad. Sci. U. S. A. 90,

4304–430857. Whitelaw, M. L., McGuire, J., Picard, D., Gustafsson, J. A., and Poellinger, L.

(1995) Proc. Natl. Acad. Sci. U. S. A. 92, 4437–444158. Pongratz, I., Mason, G. G. F., and Poellinger, L. (1992) J. Biol. Chem. 267,

13728–1373459. Wilhelmsson, A., Cuthill, S., Denis, M., Wikstrom, A. C., Gustafsson, J. A., and

Poellinger, L. (1990) EMBO J. 9, 69–7660. Whitelaw, M. L., Gustafsson, J. A., and Poellinger, L. (1994) Mol. Cell. Biol. 14,

8343–835561. Fan, C. M., Kuwana, E., Bulfone, A., Fletcher, C. F., Copeland, N. G., Jenkins,

N. A., Crews, S., Martinez, S., Puelles, L., Rubenstein, J. L. R., and Tessier-Lavigne, M. (1996) Mol. Cell. Neurosci. 7, 1–16

62. Ema, M., Morita, M., Ikawa, S., Tanaka, M., Matsuda, Y., Gotoh, O., Saijoh, Y.,Fujii, H., Hamada, H., Kikuchi, Y., and Fujii-Kuriyama, Y. (1996) Mol. Cell.Biol. 16, 5865–5875

63. Chen, H., Chrast, R., Rossier, C., Gos, A., Antonarakis, S., Kudo, J., Yamaki,A., Shindoh, N., Maeda, H., Minoshima, S., and Shimizu, N. (1995) Nat.Genet. 10, 9–10

64. Kamei, Y., Xu, L., Heinzel, T., Torchia, J., Kurokawa, R., Gloss, B., Lin, S. C.,Heyman, R. A., Rose, D. W., Glass, C. K., and Rosenfeld, M. G. (1996) Cell85, 403–414

65. Higgins, D. G., and Sharp, P. M. (1988) Gene (Amst.) 73, 237–24466. Franks, R. G., and Crews, S. T. (1994) Mech. Dev. 45, 269–27767. Jiang, B. H., Rue, E., Wang, G. L., Roe, R., and Semenza, G. L. (1996) J. Biol.

Chem. 271, 17771–1777868. McGuire, J., Coumailleau, P., Whitelaw, M. L., Gustafsson, J.-A., and

Poellinger, L. (1995) J. Biol. Chem. 270, 31353–3135769. Poon, D., Schroeder, S., Wang, C. K., Yamamoto, T., Horikoshi, M., Roeder, R.

G., and Weil, P. A. (1991) Mol. Cell. Biol. 11, 4809–482170. Sikorski, R. S., and Hieter, P. (1989) Genetics 122, 19–27

bHLH-PAS Subset That Interacts with AHR or ARNT 8593