Dioxin activates HIV-1 gene expression by an oxidative stress pathway requiring a functional cytochrome P450 CYP1A1 enzyme.

We have studied the effect of several environmental chemicals on the transient expression of a chloramphenicol acetyltransferase (cat) reporter gene linked to the promoter sequences in the long terminal repeat (LTR) of the human immunodeficiency virus type 1 (HIV-1). Aflatoxin B1, 2,3,7,8-tetrachlorodibenzo-p-dioxin (TCDD; dioxin) and benzo[a]pyrene cause a significant increases in CAT expression in mouse hepatoma Hepa-1 cells. The induction of CAT after TCDD treatment is abolished by administration of N-acetyl-L-cysteine or 2-mercaptoethanol and does not take place in a mutant cell line that lacks CYP1A1 enzymatic activity. Linker-scanning mutational analysis of transcription factor binding sites in the promoter revealed that both the NF kappa B and an adjacent aromatic hydrocarbon response element (AhRE) are required for TCDD-dependent CAT expression. In addition, mutation of the NFAT/AP-1 binding sites in the negative regulatory region of the promoter increases the magnitude of the TCDD effect. We conclude that induction of a functional CYP1A1 monooxygenase by TCDD stimulates a pathway that generates thiol-sensitive reactive oxygen intermediates which, in turn, are responsible for the TCDD-dependent activation of genes linked to the LTR. These data might provide an explanation for findings that TCDD increases infectious HIV-1 titers in experimental systems and for epidemiologic reports suggesting that exposure to aromatic hydrocarbons, such as found in cigarette smoke, is associated with an acceleration in AIDS progression.

Halogenated aromatic hydrocarbons such as 2,3 ,7,8 -tetrachlorodibenzo-p-dioxin (TCDD; dioxin) cause a profusion of apparently unrelated toxic effects in which the single common denominator is the aromatic hydrocarbon receptor-mediated transcriptional activation of the cytochrome P450 CYPIAI gene (1)(2)(3)(4)(5)(6). In humans, exposure to dioxin and various other chlorinated phenolic agents causes chloracne, a long-lasting skin disease characterized by the hyperkeratinization of follicular sebocytes (7,8). In addition, recent long-term epidemiological studies have established a link between exposure to high doses of TCDD and certain types of cancers (9,10). Dioxin is one of the strongest tumor promoters ever tested in animal model systems; it causes an elevated incidence of hepatic carcinoma and pulmonary and skin tumors (11)(12)(13) and promotes tumor formation at one-hundredth the dose of the classical tumor promoter 12-O-tetradecanoylphorbol-13-acetate (TPA) in the skin of hairless mice (1416). During rodent embryogenesis, TCDD administration also causes craniofacial abnormalities such as cleft palate and hydronephrosis (17)(18)(19)(20). Characteristic events of secondary palate formation, such as osteoblast differentiation and synthesis and mineralization of extracellular matrix, are inhibited by TCDD (21). Unlike in whole animal studies, TCDD has no toxic effect in tissue culture cells, although it causes a large elevation of intracellular calcium, which induces decreased £-adrenergic responsiveness in cardiac myocytes (22,23), and causes apoptosis of immature thymocytes (24,25). In this regard, the developing immune system is a particularly sensitive target for TCDD, with thymic atrophy being the most common pathological consequence of exposure (26). Work from our laboratory has shown that treatment of mouse hepatoma cells with polycyclic or halogenated aromatic hydrocarbons such as TCDD and benzo[a]pyrene (BaP) causes an increase in the steady-state mRNA levels of the protooncogenes c-fos, c-jun, jun-B, and jun-D and the concomitant increase of the DNAbinding activity of the transcription factor AP-1 (27). These results suggested the possibility that other transcription factors might also be activated by TCDD treatment and that genes which contain binding sites for these transcription factors in their regulatory domains might respond to TCDD or BaP treatment. We tested this hypothesis in mouse hepatoma cells by analyzing the effect of TCDD treatment on the activation of a chloramphenicol acetyltransferase (cat) reporter gene fused to the long terminal repeat (LTR) sequences of the human immunodeficiency virus-1 (HIV-1).

Materials and Methods
Aflatoxin B1 was a gift of Howard G. Shertzer, and TCDD was a gift of the Dow Chemical Company; all polycyclic aromatic hydrocarbons used were purchased from the National Cancer Institute Chemical Carcinogen Repository. The mouse cell lines used in these studies were the wild-type Hepa-1 hepatoma line (28) and its CYPlAl metabolism-deficient derivative, c37 (29)(30)(31), a variant that carries two missense mutations in the Cyplal gene, rendering the resulting enzyme nonfunctional (31). These cells were grown in a-minimal essential media supplemented with 5% fetal bovine serum.
The bacterial chloramphenicol acetyltransferase (cat) gene was used as a reporter in transient transfection experiments. The chimeric plasmid pBennCAT, carrying a fusion of the cat gene sequences to the HIV-1 U3 LTR was obtained from the National Institutes of Health AIDS Research and Reference Program. This plasmid contains approximately 500 base pairs of uncharacterized human DNA sequences (32), which were removed by standard recombinant DNA techniques, giving rise to plasmid pHIVLTRCAT. Several plasmid constructs carrying mutations in the transcription factor binding sites in the LTR were derived from the wild-type pHIVLTRCAT by linker-scanning mutagenesis (33). The sequence of the relevant portion of the U3 LTR is shown in Figure 1; the sequences that were replaced indicated by a single overline. In all cases, 10 or 30 nucleotide residues (1 or 3 helical turns) were replaced by 10 residues (one helical turn), thus preserving the relative position of the unaffected binding sites on the DNA helix. The sequence shown is from GenBank locus hivhxb2cg. Transcription factor binding sites, indicated by the overline, were determined using the transcription factor site database (TFSITES) in combination with the MAP algorithm of the DNA Analysis Package (Genetics Computer Group, Madison, Wisconsin). The double overline indicates the TATA box; transcription initiation is at residue 547. The three contiguous Spl binding sites located between residues 376 and 416 are not highlighted. chain reaction (PCR) of the wild-type plasmid using primers complementary to the flanking sequences of the site to be deleted and containing a NotI restriction enzyme site sequence at their 5' end. The PCR products were digested with NotI, ligated, and used to transfect electrocompetent E. coli DH5a. In mutant pLS1, the site replaced was the composite NFAT/AP-1 binding site located between residues 146 and 165. Mutants pLS2 (NFAT site at 200-230), pLS3 (AP-1 site at 233-243), and pLS4n (two NFiB sites at 344-374) were constructed in a similar manner. An XhoI site was used instead of NotI for mutant pLS4a (AhRE site at 378-388), and an XbaI site was used for mutant pLS5 (AP-1 site at 477-487). Double and triple mutants were constructed from the single mutants by in vitro recombinant techniques using recombination of fragments generated by appropriate restriction enzymes. All constructs were confirmed by restriction enzyme analysis and DNA sequencing.
Approximately 10 pg of the appropriate plasmid was transfected by standard calcium phosphate techniques (34,35) into semiconfluent Hepa-1 or c37 cells. In some experiments, to determine the effect of the viral transactivator Tat protein (36,37), we co-transfected the cells with plasmid pCV-1 (obtained from the NIH AIDS Research and Reference Program) that expresses HIV-1 Tat under the control of the SV40 early promoter and enhancer. We found that Tat expression increased basal CAT levels and induced CAT levels by approximately the same magnitude; all the experiments reported here were done in the absence of Tat expression. As a negative control we used pSVOCAT, containing a promoterless cat gene, and as a positive control we used pSV2CAT, carrying the cat gene under the control of the SV40 early promoter and enhancer sequences. In initial experiments, we also used TPA treatment as a positive control for HIV LTR-dependent expression. To control for variations due to differences in transfection efficiency, all cultures were co-transfected with plasmid pCMVggal (CloneTech, Palo Alto, California), which expresses the bacterial f-galactosidase gene under the control of the cytomegalovirus immediateearly enhancer and promoter. Expression of 13-galactosidase under regulation by this enhancer is independent of treatment to the cells. Twelve to 16 hr after transfection, the cells were fed with low serum (0.1%) a-minimal essential media to deplete preexisting transcription factors that respond to stimulation by serum, and 24-48 hr later the cells were treated with TCDD or other compounds or with an equivalent amount of dimethylsulfoxide vehicle. We prepared cell extracts 18 hr later by three cycles of freeze-thawing, and expression of CAT and f-galactosidase activities was determined. We measured CAT activity by the phase extraction method (38) using 0.2 pCi of 1 C]chloramphenicol (Amersham, Arlington Heights, Illinois) as the substrate. Chloramphenicol conversion to acetylated forms was 1-25%, well within the linear range [0-50% (38)] of the assay. We determined f-galactosidase activity using a kit from Promega BioTech (Madison, Wisconsin). Data were normalized for differences in transfection efficiency by determination of the relative amount of chloramphenicol converted to acetylated forms per unit of f3galactosidase. Experiments were repeated at least three times, and the values shown are the means ± SEM.

Results
We first tested the effect of aflatoxin Bl, TCDD, and several polycyclic aromatic hydrocarbons on transient CAT expression directed by the wild-type HIV-1 LTR. Mouse hepatoma Hepa-1 cells were transfected with pHIVLTRCAT; 24 hr after transfection, the cells were treated with the compounds indicated in Figure 2. Two of these compounds, aflatoxin B1 and BaP, caused approximately a fivefold increase in CAT activity over the dimethylsulfoxide control. Four others, anthracene, benzo [a] anthracene, TCDD, and 3methylcholanthrene, were next in potency, causing a two-to threefold increase in CAT activity. The effects of the last two, dimethylbenzanthracene (DMBA) and dibenzo[a,h]anthracene (DBA), were not significantly different from the control (Fig. 2). In initial control experiments, we estimated that the effect of BaP on CAT expression was approximately 50% of maximal activation obtained with TPA (data not shown).
The HIV-1 LTR is a regulatory domain approximately 500 bases long that contains several sequence motifs recognized by cellular transcription factors. These recognition motifs include binding sites for AP-1, AP-2, AP-3, AP-4, NFAT-1, USF, URS, NFiB (39,40), as well as the TAR sequence, recognized by the viral Tat protein (3X. The initial rate of proviral transcription is determined by interactions between these transcription factors and their cognate sequences in the LTR (36,37,(39)(40)(41)(42). Since TCDD induces AP-1 activity (27), one possible explanation for the stimulation of CAT activity by TCDD was that it resulted from TCDD-induced increases in AP-1. Alternatively, CAT stimulation could be due to activation of transcription factors other than AP-1. To address this question and to analyze which transcription factor binding sites were responsible for CAT induction by TCDD, we prepared a collection of single, double, and triple mutant derivatives of the reporter plasmid and measured CAT expression directed by these mutants in transient expression assays. Mutation of the proximal AP-1 in LS5 had no effect on basal expression levels but caused an increase in TCDD-dependent CAT expression, whereas mutation of the distal NFAT and AP-1 sites in LS 1, LS2, and LS3 had a negligible effect on both basal and TCDD-induced CAT expression (Fig.  3). Mutation of the NFicB sites in LS4n or of the cryptic AhRE site in LS4a almost completely abolished basal and TCDDstimulated CAT expression (Fig. 3). Double mutations caused diverse effects: LS12 abolished TCDD induction and increased basal expression slightly, whereas LS 15 and LS35 showed both elevated basal expression levels and absence of TCDD stimulation; LS23 and LS25 had low basal levels and a TCDD stimulation factor of 6to 7-fold ( Fig. 3 and Table 1). The LS123 triple mutant exhibited extremely elevated basal as well as TCDD-stimulated Volume 103, Number Figure 3. Effect of TCDD on CAT expression directed by linker-scanning mutants. The dial shows the approximate position of the mutated sites, with the individual mutations denotE cle (see Figure 1 for the actual coordinates and the text for a complete description of eac Relative CAT activity values were determined as indicated in Figure 2. levels, but the extent of induction was not significantly different than in the wild-type ( Fig. 3 and Table 1). These results indicate that the NFicB sites mutated in LS4n are required for basal and TCDD-stimulated CAT expression. Furthermore, a previously unrecognized Ah receptor bin canonical GCGTG AhRE sit 380 mutated in LS4a, was als4 essential for expression in he In addition, the region betv 115 and 255 that contains th The extent of more detail in the next section.
The involvement of an AhRE site on CAT expression suggested the possibility that the Ah receptor and a TCDDinducible cytochrome P450 CYPlAl enzyme might participate in the stimula- 10 12 tion of CAT expression observed after with increasing doses of either compound.
At a dose of 1 mM NAC or 20 1iM 2-mercaptoethanol, the effect of TCDD was completely abolished (Fig. 5), suggesting that, indeed, GYPIlAl-dependent oxidative stress might be responsible for the effect of TCDD on LTR-directed CAT expression.

Discussion
The results that we present in this article indicate that TCDD, aflatoxin B1, and several polycyclic aromatic hydrocarbons (PAHs) can significantly activate the expression of genes linked to the LTR sequences of HIV-1. The magnitude of the activation appears to be different for the various compounds tested. In the case of TCDD, stimulated values are significantly higher than control values, although they do not usually exceed them by more than 2.5to 3-fold. This stimulation is in agreement with observations by others that TCDD can cause an increase of infectious HIV-1 titers in experimental systems (43,44). As for PAHs, the highest levels of CAT activation that we observed were a result of BaP treatment, a finding that may provide a possible molecular explanation for the observation that cigarette smoking accelerates the progression of AIDS (45--48). It is, of course, likely that the effect of cigarette smoke on AIDS progression results from a combination of many different causes, of which gene activation by BaP is only one. Surprisingly, DBA had no effect on CAT expression, a finding that we cannot explain at present. Using linker-scanning mutational analysis, we have identified several domains of the HIV-1 LTR responsible for basal as well as TCDD-stimulated CAT expression. We find that expression directed by the HIV-1 LTR is high in mouse hepatoma cells, in agreement with previous observations in human hepatoma cell lines (49)(50)(51), this suggests that the liver may be a primary virus reservoir. Mutation of the NFICB binding sites eliminates CAT expression, confirming the absolute requirement for NFiB. NFiCB, however, is not the only transcription factor necessary for expression; we have uncovered an Ah receptor response element containing the canonical AhRE sequence GCGTG, which is also essential for basal expression. In addition, this site participates in the effect of TCDD on CAT activation because its mutation in pLS4a reduces drastically the fold induction by TCDD (Table 1). This AhRE site is embedded within the first of three adjacent Spl sites, which have been shown to interact cooperatively with NFPcB in HIV enhancer activation (52). It could be argued that SpI, and not the Ah receptor, was the transcription factor responsible for the loss of activity of pLS4a because both binding sites would be equally affected by the mutation. This possibility is unlikely because mutations in just one of the three Spi binding sites have little or no effect on HIV enhancer expression (40); this suggests that the Ah receptor, and not Spi, is the relevant transcription factor whose binding and subsequent activity are affected by the LS4a mutation. As shown by the mutagenesis analysis, both NFicB and Ah receptor binding sites are responsible for the basal expression levels, and it is possible that both transcription factors function in synergy.
The NFAT and AP-1I binding sites in mutants LS 1, LS2, and LS3, clustered in the negative reguatory region of the LTR, do not show a major effect on expression when altered individually. In double mutants that include LS 1, as well as in the triple mutant 1.5123, basal level of expression is elevated, suggesting that the NFAT/AP-1 site at 146-165 is responsible for downregulating the basal level of expression. This is in agreement with earlier findings that this region of HIV-1 contains negative regulatory elements (39,40). As shown in Table 1, these double mutants and the triple mutant L5123 also show a low level of induction by Volume 103, Number 4, March 199536 I~~~~~~* TCDD, suggesting that the same sites that downregulate the basal level of expression may respond to TCDD activation. In contrast, mutations at the LS2 site, and possibly at the LS5 site, cause a significant increase in the fold induction by TCDD that reach levels double those of the wild-type (Table  1); this suggests that the role of the LS2 site is antagonistic to that of the LSI site, having a dampening effect on TCDD induction when not modified.
Interpretation of these results must take into account recent findings regarding the properties of transcription factor AP-1: 1) different combinations of fos and jun family members have very different effects on the same promoter (53); 2) different promoters respond differently to the same combination of fos and jun family members (53); and 3) fos and jun are integral components of the NFAT complex (54). Consequently, sites LS1, LS2, LS3, and LS5 may have antagonistic roles due to conflicting effects of free AP-1 and ofAP-1 in NFAT complexes on different promoter sequences. The overall transcriptional effect of this combination of antagonistic sites would be rather unpredictable. Within this context, the outcome of TCDD exposure is likely to result from a combination of two opposing effects; on one hand, activation of expression may take place by means of the Ah receptor, the AhRE site, and a particular set of fos/jun members with positive effects on LTR expression. On the other hand, dampening of this induction of expression may occur by activation of other fos/jun members with negative regulatory functions. TCDD toxicity has been proposed to result from epoxides and other derivatives of arachidonic acid metabolism catalyzed by TCDD-induced cytochrome P450 enzymes (23,55,56). In agreement with this hypothesis, we find that stimulation of CAT expression by TCDD is absent in the variant cell line c37 that lacks cytochrome P450 CYPlAl activity, strongly suggesting that the effect of TCDD is mediated by the monooxygenase activity of CYPlAl. This activity might generate arachidonate metabolites responsible for the elevation of the pro-oxidant status of the cell, and indeed several cytochromes P450, including the TCDD-inducible CYPlAl and CYP1A2 enzymes (57)(58)(59) and others (60,61), possess arachidonic acid epoxygenase activity. Our experiments, although not directly aimed at the identification of possible mediators, show that NAC and 2mercaptoethanol eliminate the effect of TCDD, indicating that, as shown for NFKB activation (62)(63)(64), oxidative stress caused by thiol-sensitive reactive oxygen species is likely to be involved in the TCDD-dependent activation events.
In conclusion, our data are consistent with a signal transduction mechanism that includes at least two different TCDDdependent pathways. On one hand, activation of the Ah receptor triggers expression mediated by the AhRE site present in the LTR. On the other hand, TCDD induction of a functional CYPlA1 monooxygenase stimulates generation of thiol-sensitive reactive oxygen species, which in turn activate transcription factors operative in LTRdirected expression. We find that, in addition to TCDD, several other toxic environmental chemicals can activate expression of the HIV-1 promoter-enhancer sequences, underscoring the importance that exposure to these compounds might have in the progression ofAIDS. This years topic is "Environmental Lung Disease: Exposures & Mechanisms." We will explore the common ground shared by researchers who study the basic mechanisms of occupational and environmental lung disease and who study the impact of environmental exposure on human populations. Topics will include cellular, molecular, immunologic and genetic mechanisms involved in the response to environmental and occupational toxicants; and the clinical and epidemiologic relationship of inhalational exposure to lung diseases of the airway and interstitium, including asthma, fibrosis, granulomatosis, and malignancy.