Down-regulation of Major Histocompatibility ComplexQ1 b Gene Expression by 2,3,7,8-Tetrachlorodibenzo-p-dioxin*

We analyzed mouse hepatoma cells using differential display to discover new genes that respond to the environmental contaminant 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). We identified a class I major histocompatibility complex (MHC) gene, which we designated as MHC Q1 b , whose expression decreases in the presence of TCDD. TCDD-induced down-regulation of MHCQ1 b requires both the aromatic hydrocarbon receptor and the aromatic hydrocarbon receptor nuclear translocator, transcription factors that up-regulate other genes in response to TCDD. Down-regulation of MHC Q1 b by TCDD appears to involve both transcriptional and post-transcriptional regulatory events; the post-transcriptional destabilization of MHCQ1 b mRNA is probably a secondary response to TCDD. Our findings reveal new mechanistic aspects of gene regulation by TCDD. In addition, our observations suggest a mechanism that might account for some of TCDD’s immunotoxic effects.

We analyzed mouse hepatoma cells using differential display to discover new genes that respond to the environmental contaminant 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD). We identified a class I major histocompatibility complex (MHC) gene, which we designated as MHC Q1 b , whose expression decreases in the presence of TCDD. TCDD-induced down-regulation of MHC Q1 b requires both the aromatic hydrocarbon receptor and the aromatic hydrocarbon receptor nuclear translocator, transcription factors that up-regulate other genes in response to TCDD. Down-regulation of MHC Q1 b by TCDD appears to involve both transcriptional and posttranscriptional regulatory events; the post-transcriptional destabilization of MHC Q1 b mRNA is probably a secondary response to TCDD. Our findings reveal new mechanistic aspects of gene regulation by TCDD. In addition, our observations suggest a mechanism that might account for some of TCDD's immunotoxic effects.
2,3,7,8-Tetrachlorodibenzo-p-dioxin (TCDD) 1 represents the prototype for a class of structurally related halogenated aromatic hydrocarbons, including polychlorinated dibenzo-p-dioxins, dibenzofurans, and biphenyls. Many such compounds are widespread environmental contaminants, produce similar patterns of toxicity, and appear to share the same receptor-mediated mechanism of action. Because TCDD is the most potent, it has been studied much more extensively than other compounds. TCDD is generated during combustion processes, enters the atmosphere, contaminates the food chain, and persists in the environment, because it is resistant to biological degradation. In animals, TCDD elicits numerous adaptive and/or adverse effects, including enzyme induction, epithelial hyperplasia, teratogenesis, tumor promotion, liver toxicity, thymic atrophy, and immunosuppression. In humans, the most well documented adverse response to TCDD is chloracne; there is also concern that TCDD produces cancer, birth defects, and immunotoxicity (1)(2)(3). The molecular mechanism of TCDD action probably involves alterations in gene expression, which are mediated by the aromatic hydrocarbon receptor (AhR). AhR is a cytoplasmic, basic helix-loop-helix protein that binds TCDD saturably and with high affinity, thereby activating signaling pathways that modulate gene expression (4,5).
The best understood AhR-regulated response to TCDD is the induction of CYP1A1 gene transcription (5). CYP1A1 encodes the microsomal enzyme cytochrome P4501A1, which catalyzes the oxygenation of lipophilic aromatic hydrocarbons during their metabolic processing to water-soluble derivatives (6). TCDD binds to AhR in the cytoplasm; subsequently, the liganded AhR enters the nucleus and heterodimerizes with a second basic helix-loop-helix protein known as the AhR nuclear translocator (Arnt). The AhR/Arnt heterodimer functions as a liganddependent transcription factor; it binds to a specific DNA sequence, termed a xenobiotic-responsive element, within an enhancer upstream of the CYP1A1 gene (7,8). This protein-DNA interaction is associated with disruption of nucleosomes, occupancy of the CYP1A1 promoter by transcription factors, and activation of gene expression (9 -11). In comparison with activation of transcription, we know relatively little about the down-regulation of gene expression by TCDD.
One way to increase our understanding of TCDD action is to analyze additional TCDD-responsive genes. To address this issue experimentally, we used differential display, a versatile method that can identify both up-and down-regulated genes (12,13). We studied mouse hepatoma cells, because the availability of AhR-defective and Arnt-defective mutants permits both genetic and biochemical analyses of the regulatory mechanism (14,15). Here, we show that TCDD down-regulates the expression of a class I major histocompatibility complex (MHC) gene, which we designate as MHC Q1 b , and we examine the mechanism of the down-regulation.
Differential Display-We used a reverse transcription-PCR-based differential display approach (12,13). Experiments were carried out using the RNAimage kit according to the manufacturer's instructions (GenHunter). Briefly, total RNA was isolated from untreated or TCDDtreated (1 nM, 16 h) wild-type cells using a Qiagen total RNA isolation kit and was reverse-transcribed using as primers one of the three different one base-anchored oligo(dT) 11 primers, H-TM (where M may be A, C, or G), containing a HindIII site. cDNAs were amplified by PCR in the presence of [ 33 P]dATP using the H-TM primer and a 13-mer arbitrary primer, H-AP, containing a HindIII site for subcloning. PCR * These studies were supported by NCI, National Institutes of Health (NIH), Grant CA 53887 (to J. P. W.) and NIEHS, NIH, Grant NRSA ES 05679 (to Q. M.). The costs of publication of this article were defrayed in part by the payment of page charges. This article must therefore be hereby marked "advertisement" in accordance with 18 U.S.C. Section 1734 solely to indicate this fact.
‡ To whom correspondence should be addressed. Tel.: 415-723-8233; Fax: 415-725-2952. 1 The abbreviations used are: TCDD, 2,3,7,8-tetrachlorodibenzo-pdioxin; AhR, aromatic hydrocarbon receptor; Arnt, Ah receptor nuclear translocator; AURE, adenylate-uridylate-rich element; DIG, digoxigenin; MHC, major histocompatibility complex; PCR, polymerase chain reaction; kb, kilobase pair(s). products were separated on a 6% DNA sequencing gel and detected by autoradiography. Bands exhibiting differential expression were excised from the dried gel, and DNAs on the gel slice were eluted and amplified by PCR using the same set of primers. The PCR products were subcloned into a TA cloning vector, pCR II or pCR 2.1 (Invitrogen), and sequenced using a Sequenase kit. Taq polymerase was used in all PCR steps. The conditions for reverse transcription and PCR were as recommended by the manufacturer (GenHunter).
cDNA Library Screening-To obtain full-length cDNA, we screened a Hepa 1c1c7 cell cDNA library (17), using clone C38 from the differential display (see "Results"). A DIG-labeled DNA probe was synthesized according to the manufacturer's instructions (Boehringer Mannheim), using clone 38 as a template. The labeled probe was used to screen 1 ϫ 10 6 phage plaques. Positive clones, contained within the pBK-CMV phagemid, were excised in vivo from the ZAP Express vector using the ExAssist-SOLR system, as recommended by the manufacturer (Stratagene). The excised phagemids containing the positive clones were sequenced by the dideoxy method using a Sequenase kit.
RNA Analyses-For Northern blotting, a cDNA fragment encoding exons 5 and 6 of MHC Q1 b was generated by PCR, subcloned into pBluescript, and used as a template for riboprobe synthesis. Riboprobes were synthesized in the presence of DIG-UTP using a DIG RNA-labeling kit (Boehringer Mannheim). Total RNA was isolated from cells using a Qiagen total RNA isolation kit. RNA (10 g) was electrophoresed on a 1% agarose-formaldehyde gel and transferred to a nitrocellulose membrane (Schleicher & Schuell). After cross-linking, the membrane was hybridized with DIG-labeled riboprobe at 68°C for 16 h; signals were detected by chemiluminescence using a DIG RNA labeling and detection kit with CDP Star as substrate (Boehringer Mannheim).
RNase Protection-cDNAs in pCR II or pCR 2.1 were used as DNA templates to synthesize 32 P-labeled riboprobes. A cDNA fragment encoding exon 2 of MHC H-2K b , H-2D b , Q1 b , or Q10 b was generated by PCR, subcloned into pBluescript, sequenced, and used for riboprobe synthesis. The DNA templates for MHC H-2K b (18), H-2D b (19), and Q10 b (20) were gifts from Dr. Larry R. Pease (Mayo Clinic). The DNA template for MHC Q1 b was obtained from cDNA library screening. The transcription reaction was carried out in vitro in the presence of [␣-32 P]UTP using an Ambion transcription kit. Total RNA (5 g) was hybridized with a riboprobe at 50°C for 16 h. tRNA was used as a negative control. Single-stranded RNA was digested with RNase A and RNase T1. Protected fragments were separated on a 6% polyacrylamide/urea gel and detected by autoradiography. A 1-kb DNA ladder was phosphorylated and electrophoresed in a parallel lane to estimate fragment size.
Nuclear Run-on-Nuclear run-on experiments were performed as described by Ausubel et al. (21) with modifications. Cells were lysed in 10 mM Tris-Cl, pH 7.4, 10 mM NaCl, 3 mM MgCl 2 , and 0.5% Nonidet P-40. Nuclei were isolated by centrifugation for 5 min at 500 ϫ g. In vitro transcription was carried out at 30°C for 30 min in a transcription buffer containing 5 mM Tris-Cl, pH 8.0, 2.5 mM MgCl 2 , 0.15 M KCl, 0.1 mM EDTA, 0.5 mM dithiothreitol, 0.5 mM ATP, 0.5 mM CTP, 0.5 mM GTP, 0.175 mM UTP, and 0.325 mM DIG-UTP (Boehringer Mannheim). DIG-labeled RNA was purified using a Qiagen total RNA isolation kit. cDNAs immobilized on a nitrocellulose membrane were hybridized with DIG-labeled, purified RNA at 42°C for 16 h. The membrane was washed, and the signal was detected by chemiluminescence according to the manufacturer's recommendation (Boehringer Mannheim).

Identification of TCDD-regulated Genes by Differential
Display-We employed a differential display technique to identify TCDD-responsive genes in Hepa 1c1c7 cells. As a positive control, we identified CYP1A1 (clone C17, Fig. 1), a gene that is known to be up-regulated by TCDD. In addition, we identified additional up-regulated genes and a down-regulated gene. We confirmed that TCDD regulates these genes using slot blot analyses and/or RNase protection studies (data not shown). These genes were cloned and sequenced. A GenBank TM search revealed that none of these cDNA sequences had been reported previously. We chose to analyze the down-regulated gene (clone C38, Fig. 1) because of the likelihood that such studies would reveal new mechanistic aspects of TCDD action.
Identification of a Full-length cDNA for Clone C38 -To obtain full-length cDNA, we used clone C38 as a probe to screen a mouse hepatoma (Hepa 1c1c7) cDNA library (17). cDNAs that contained the C38 sequence were sequenced; a GenBank TM search indicated that the cDNAs are nearly identical to the MHC Q1 k gene (mouse C3H strain) (22), a nonclassical, class I MHC gene (see "Discussion"), in all eight predicted exons ( Fig.  2). At the 3Ј-end, all clones contain a poly(A) tail and a 528nucleotide untranslated sequence; this untranslated region contains the clone C38 sequence, but it is not present in the MHC Q1 k sequence. These observations explain why we did not find a sequence corresponding to C38 in GenBank TM . We have designated the down-regulated gene as MHC Q1 b (mouse C57BL/6 strain).
Typically, mouse class I MHC genes contain eight exons, each of which encodes a separate protein domain (23). Exon 1 encodes a leader sequence. Exons 2 and 3 encode the extracellular ␣1 and ␣2 domains, which are involved in peptide binding. Exon 4 encodes the ␣3 domain, which interacts with ␤ 2microglobulin. Exon 5 encodes the transmembrane domain. Exons 6 -8 encode the cytoplasmic domain. The sequences of seven cDNA clones for MHC Q1 b imply that it exhibits substantial variation in RNA splicing, compared with MHC Q1 k (22). For example, 1) one clone contains intron 1 (Fig. 3A), which could encode an open reading frame and could be in frame between exon 1 and exon 2; 2) four clones contain introns 5, 6, and/or 7 in various combinations (Fig. 3, B-D); 3) one clone lacks exon 3; 4) three clones are spliced from exon 5 to exon 8, resulting in deletion of exons 6 and 7. Only two of seven clones are normally spliced from exon 5 to exon 6. All clones contain exons 4 and 5, which encode the ␣3 domain and the transmembrane domain, respectively; these latter findings suggest that MHC Q1 b may interact with ␤ 2 -microglobulin and may be membrane-anchored.
Comparison of DNA sequences in the predicted coding regions of Q1 b and Q1 k reveals four nucleotide mismatches, which produce one amino acid change; in addition, one nucleotide insertion in exon 7 of Q1 b produces an open reading frameshift and the addition of five amino acids (Fig. 2). These sequence data suggest that MHC Q1 genes are relatively nonpolymorphic, a characteristic that is typical of nonclassical class I MHC genes (20).
Expression of MHC Q1 b mRNA-We performed Northern blots to determine the MHC Q1 b transcript size. If the majority of transcripts contain eight exons plus the 528-nucleotide untranslated sequence, the transcript size should be ϳ2 kb. If the majority of transcripts contain unspliced introns identified in the cDNA clones, the transcript size should be ϳ2.7 kb. To identify an MHC Q1 b -specific probe, we synthesized several probes by PCR using cDNA clones as templates, and we analyzed RNA from uninduced and TCDD-induced wild-type cells. Northern blots probed with exon 2, 3, or 4 showed two transcripts, about 1.6 and 2.7 kb in size. Expression of the 2.7-kb transcript was down-regulated by TCDD; in contrast, expres-FIG. 1. Identification of TCDD-regulated genes by differential display. Wild-type cells were treated with Me 2 SO (Ϫ) as control or 1 nM TCDD (ϩ) for 16 h. Total RNA was used for differential display. The arrows indicate the differentially displayed mRNAs. sion of the 1.6-kb transcript was not affected by TCDD treatment (data not shown). Clone C38, exons 5 and 6, intron 5, and intron 6 hybridized only with the 2.7-kb transcript (data not shown). Because down-regulation of the 2.7-kb transcript by TCDD is not accompanied by the appearance of other transcripts or an increase in the 1.6-kb transcript, we infer that the TCDD-induced decrease in the 2.7-kb transcript is not due to a change in the pattern of RNA splicing. Our results indicate that Q1 b mRNA is differentially spliced, that the 2.7-kb transcript represents most MHC Q1 b mRNA, which contains unspliced introns, and that MHC Q1 b mRNA is down-regulated by TCDD. The 1.6-kb transcript may represent cross-hybridization of exons 2-4 to H-2D b mRNA, which is ϳ1.6 kb in size (24,25) and is not regulated by TCDD, as described later. A probe containing exons 5 and 6 generated a strong signal with a low background; therefore, we used this probe for the remaining analyses.
Analyses of MHC Q1 b mRNA from wild-type cells reveal that TCDD down-regulates MHC Q1 b mRNA in a time-dependent manner (Fig. 4). Assuming a first order degradation rate, the half-life of MHC Q1 b mRNA after TCDD treatment is about 7 h. In addition, the down-regulation of MHC Q1 b mRNA by TCDD is dose-dependent (Fig. 5). The EC 50 is about 30 pM, a value similar to that for the induction of CYP1A1 mRNA by TCDD, a response that is mediated by the Ah receptor.
AhR Dependence and Arnt Dependence of MHC Q1 b Downregulation by TCDD-AhR and Arnt mediate most, if not all, responses to TCDD (5,7,8). Therefore, we asked whether down-regulation of MHC Q1 b expression by TCDD requires AhR and/or Arnt. We employed variant cells defective in either AhR or Arnt to address this question.
Our findings indicate that TCDD fails to down-regulate MHC Q1 b in Arnt-defective cells; down-regulation is restored by reconstitution of Arnt-defective cells with wild-type Arnt cDNA but not by reconstitution with a mutant Arnt cDNA (Fig.  6). These results indicate that down-regulation of MHC Q1 b by TCDD requires Arnt.
For reasons that we do not understand, AhR-defective cells do not express MHC Q1 b mRNA (data not shown). Therefore, we cannot use AhR-defective cells to study MHC Q1 b gene regulation. Instead, we used wild-type cells in which we expressed a mutant form of AhR, R39A, which has a dominant negative effect on TCDD-inducible gene expression (16). The mutant AhR can heterodimerize with Arnt to generate a nonfunctional heterodimer, which cannot bind to DNA; presumably, the mutant AhR exerts its dominant negative effect by competing with wild-type AhR. Expression of the dominant negative AhR mutant in wild-type cells reduced the extent to which TCDD down-regulates MHC Q1 b (Fig. 7); this finding implies that down-regulation of MHC Q1 b by TCDD requires AhR. This conclusion is consistent with our observation that the EC 50 for MHC Q1 b down-regulation is similar to that for a known AhR-dependent response, CYP1A1 induction. In addition, our conclusion is consistent with the Arnt dependence of MHC Q1 b down-regulation, because AhR-mediated responses also require Arnt.
Transcriptional and Post-transcriptional Regulation of MHC Q1 b Gene Expression by TCDD-Down-regulation of MHC Q1 b by TCDD could be due to a decrease in the rate of RNA synthesis, to an increase in the rate of RNA degradation, or to both. To determine whether TCDD regulates MHC Q1 b at the level of transcription, we performed nuclear run-on experiments. Multiple experiments (see Fig. 8 for example) reveal that TCDD consistently decreases the rate of MHC Q1 b transcription; however, we observed substantial variation in the extent of the decrease, ranging from 2-to 10-fold. This finding suggests that the down-regulation of MHC Q1 b by TCDD occurs, at least in part, at the transcriptional level.
We used actinomycin D to determine whether TCDD also has a post-transcriptional effect and influences the degradation rate of MHC Q1 b mRNA. When cells are exposed to actinomycin D at a concentration that inhibits transcription by Ͼ95% (26), MHC Q1 b mRNA remains elevated for at least 12 h (Fig.  9). This finding implies that, in the absence of TCDD, MHC Q1 b mRNA has a relatively long half-life, because its concentration does not fall when transcription is blocked. Therefore, because MHC Q1 b mRNA concentration falls in the presence of TCDD, we infer that TCDD must increase the rate of MHC Q1 b mRNA degradation. We note that actinomycin D blocks the TCDDinduced decrease in MHC Q1 b mRNA (Fig. 9). This result implies that TCDD acts indirectly to increase MHC Q1 b mRNA degradation, because the effect requires ongoing RNA synthesis. Taken together, our studies imply that TCDD influences the regulation of MHC Q1 b gene expression at both transcriptional and post-transcriptional levels.  (Fig. 10). We did not detect H-2K b and Q10 b mRNAs in these cells. These results indicate that TCDD does not down-regulate all class I MHC genes; its effect may be specific for MHC Q1 b gene expression. DISCUSSION To generate a broader perspective on the biological effects of TCDD and to uncover novel mechanistic aspects of TCDD action, we used differential display to find new dioxin-responsive genes in mouse hepatoma cells. We isolated full-length cDNA for the TCDD-responsive MHC Q1 b gene, and we analyzed the mechanism of its down-regulation.

Specificity of TCDD's Effect on MHC Q1 b Gene Expression-
Sequence analyses of MHC Q1 b cDNA clones reveal high homology to the MHC Q1 k gene. The analyses also imply the existence of several splicing variants. However, none of the MHC Q1 b clones is spliced as predicted for MHC Q1 k (22). Three of seven clones are spliced from exon 5 to exon 8, which deletes the protein's cytoplasmic domain. Two clones contain intron 5, which produces an abnormal shortening of the cytoplasmic domain. Multiple spliced transcripts have been observed in studies of other nonclassical class I MHC mRNAs (27)(28)(29). Our observations extend these previous findings and suggest that multiple mRNA splicing may be common in nonclassical class I MHC gene expression.
Classical class I MHC proteins, which are encoded by the H-2 K, D, and L regions of mouse chromosome 17 (see Ref. 40 and references therein) present foreign peptides primarily to cytotoxic T cells, which lyse the cells that harbor the antigen (such as a viral peptide). In contrast, the function of nonclassical class I MHC proteins, which are encoded by the Q and T regions, is not well understood; these proteins may have di- verse functions. For example, the product of the MHC Q10 gene is secreted (30,31), whereas the product of the MHC Q5 gene is cytoplasmic (32). Therefore, MHC Q10 and Q5 are unlikely to be involved in antigen presentation to cells of the immune system. On the other hand, the nonclassical class I MHC T10 and T22 gene products may participate in presenting antigen to T cells with ␥␦ receptors (33,34). Our findings reveal that each MHC Q1 b cDNA encodes the ␣3 domain and the transmembrane domain; these observations imply that the MHC Q1 b gene product could interact with ␤2-microglobulin and could be a membrane protein. Therefore, we envision that MHC Q1 b could function in antigen presentation; however, this hypothesis remains to be tested. We used RT-PCR to analyze MHC Q1 b gene expression in tissues, and we detected Q1 b mRNA in mouse liver, testis, kidney, lung, and spleen. 2 Therefore, MHC Q1 b might function in numerous tissues.
TCDD elicits a variety of immunological effects in experimental animals, including suppression of both T cell-mediated immunity and humoral immunity (3). Down-regulation of a class I MHC gene(s) represents a direct mechanism by which TCDD could suppress T cell-mediated immunity, because it might adversely affect antigen presentation. For example, mice exposed to TCDD exhibit enhanced susceptibility to viral infection (35). We speculate that down-regulation of MHC gene expression by TCDD could hinder the presentation of viral peptides to cytotoxic T cells, thereby enhancing survival of the virus and increasing the susceptibility of the host to its adverse effects.
The mechanism by which TCDD down-regulates gene expression has not been studied in detail. Our findings indicate that down-regulation of MHC Q1 b mRNA is an AhR/Arntmediated response and occurs at both transcriptional and posttranscriptional levels. We can envision several possible mechanisms for transcriptional down-regulation: 1) the AhR/Arnt heterodimer could interact directly with a silencer-like element(s) (36) on the MHC Q1 b gene; 2) AhR/Arnt could disrupt the DNA binding of a positive regulatory factor(s) by interacting at an overlapping AhR/Arnt recognition sequence (37); 3) AhR/Arnt could interact directly with another transcription factor(s), thereby squelching constitutive expression of MHC Q1 b ; 4) AhR/Arnt could induce a gene whose product downregulates MHC Q1 b transcription (38). The identification of MHC Q1 b transcriptional regulatory elements and analyses of protein-DNA interactions at these elements may enable us to test these hypotheses in the future.
Our findings imply that post-transcriptional regulation must be important in TCDD-induced down-regulation of MHC Q1 b gene expression, because MHC Q1 b mRNA is stable in the absence of ongoing RNA synthesis. We note that the 3Ј-untranslated region of MHC Q1 b contains four adenylate-uridylate-rich elements (AUREs), AUUUA. Such elements can function as mRNA-destabilizing signals by an unknown mechanism (39). Therefore, we speculate that the AUREs destabilize MHC Q1 b mRNA in TCDD-treated cells. Our observations also suggest that destabilization of MHC Q1 b mRNA is a secondary response to TCDD. Therefore, as a working hypothesis, we envision that TCDD induces the transcription of a gene whose product destabilizes MHC Q1 b mRNA by recognizing its AUREs. Such post-transcriptional control is a novel aspect of TCDD action; we hypothesize that it may affect the expression of additional TCDD-responsive genes, thereby contributing to the diversity of dioxin's biological effects.