To identify human CAR splicing variants, we cloned the cDNA of CAR by PCR using oligo(dT)-primed cDNA of an individual liver and of differentiated Caco-2 cells as templates. The intestinal Caco-2 cells show an induced CAR mRNA expression during enterocytic differentiation (see Fig. 7A). Analyzing the sequences of 23 clones containing the complete open reading frame, we identified the originally published CAR cDNA sequence 4 here referred to as SV1 (8 clones), and the splicing variants SV2 (11 clones), SV4 (3 clones) and SV6 (1 clone) which arise from three different alternative splicing events (Fig. 1A), all resulting in ligand binding domain variant CAR proteins. Two of the splicing events were generated using alternative intronic 3'splice sites, which result in the addition of 12 bp or 15 bp to the 5'end of exon 7 or exon 8, respectively. The encoded protein variants are characterized by the in-frame insertion of 4 amino acids (VSPT) or 5 amino acids (APYLT), respectively. The third splicing event, skipping of exon 7, results in an in-frame deletion of 39 amino acids. While this work was in progress, splicing variants rising from these three splicing events have been described 16. By sequencing 10 cDNA clones encompassing exons 8–9, we identified a fourth alternative splicing event, which uses an alternative 3'splice site within exon 9 leading to the out-of-frame deletion of the first 76 bp of the exon. Consequently, the last 42 amino acids of CAR are replaced by 7 new amino acids in proteins encoded by splicing variants containing this out-of-frame deletion (Fig. 1B). If all possible combinations of the four individual alternative splicing events are taken into account, 12 different splicing variants of CAR can be expected which we named SV1-SV12 (Fig. 1A).

To analyze the impact of the alterations in the LBD generated by the different splicing events on CAR function, we constructed CAR expression plasmids encoding variant proteins which harbor single changes in the LBD: CAR(SV2), CAR(SV3), CAR(SV4), CAR(SV5) (Fig. 2A). In addition we constructed an expression plasmid encoding CAR(SV6), as we found this particular combination of splicing events in a full length open reading frame cDNA clone, while we did not find any clone containing only the insertion of 12 bp in front of exon 7.

We first investigated whether the intracellular distribution of the variants was altered, compared to reference CAR(SV1). Transient expression of reference CAR(SV1) in COS1 cells demonstrated that the protein was accumulating slightly stronger in the cytoplasm than in the nucleus (Fig. 2B). All CAR protein variants were distributed similarly. The apparent sizes of the variants were in good agreement with the calculated molecular weight of approximately 40 kDa for SV1, SV2, SV3 and SV6 and 35 kDa for SV4 and SV5 (Fig. 2B). Furthermore all protein variants were efficiently and stably expressed in COS1 cells. There was no indication of degradation of particular isoforms. In conclusion these findings indicate that the changes within the LBD of the variant CAR proteins neither influence protein stability nor intracellular distribution.

To transactivate the expression of its target genes, CAR has to bind to its response elements as a heterodimer with RXRα. Therefore, we analyzed heterodimerization in a mammalian two hybrid assay. Fig. 3A shows that the LBD of reference CAR(SV1) and of CAR(SV3) did interact constitutively with the LBD of RXRα. Remarkably, interaction between RXRα and CAR(SV1) was about 10-fold stronger than interaction between RXRα and CAR(SV3). The LBD of all the other variants failed to interact with RXRα. Treatment with the human CAR-specific agonist CITCO 10 did not influence the interaction of any CAR isoform with RXRα (data not shown).

To investigate DNA binding of CAR variant proteins to response elements in the XREM of CYP3A4 17 and in the -7.8 kb enhancer of MDR1 (Burk et al., to be published elsewhere), we performed electrophoretic mobility shift assays using in vitro translated RXRα and CAR isoforms. Fig. 3B shows that all CAR protein isoforms were expressed at comparable levels in vitro. However, only reference CAR(SV1) was able to bind to the CAR-response elements DR3 of CYP3A4-XREM (Fig. 3C) and DR4(I) of MDR1 -7.8 kb enhancer (data not shown) as a heterodimer with RXRα.

To further analyze the functional activity of the CAR LBD variant proteins, we performed transient co-transfection assays with enhancer/promoter reporter gene constructs of CYP2B6, CYP3A4 and MDR1 in COS1 cells. Co-transfection of reference CAR(SV1) expression plasmid resulted in 11-fold activation of CYP2B6 reporter, 7-fold activation of CYP3A4 reporter and 12-fold activation of MDR1 reporter, thus demonstrating constitutive transcriptional activity of reference CAR(SV1) in COS1 cells (Fig. 4). In contrast, CAR protein variants CAR(SV2), CAR(SV4), CAR(SV5) and CAR(SV6) were not able to transactivate any of the reporter gene constructs significantly. However, CAR(SV3) significantly transactivated the CYP2B6 and MDR1 reporter genes 3-fold and 7-fold, respectively, whereas the CYP3A4 reporter gene was not significantly activated (Fig. 4).

CAR demonstrates constitutive interaction with coactivators, which is further enhanced by ligand binding 9. We analyzed the interaction of CAR isoforms with different coactivators in mammalian two hybrid experiments (Fig. 5). Reference CAR(SV1) showed a strong constitutive interaction with the coactivators tested. The strongest interaction was observed with DRIP205 (100-fold). None of the variant proteins interacted constitutively with coactivators of the p160 family (SRC-1, TIF-2 and ACTR) in this assay. A weak interaction of CAR(SV3) with DRIP205 was observed, whereas the other isoforms did not interact with DRIP205. As expected, the human CAR-specific ligand CITCO enhanced interaction of reference CAR(SV1) with each of the coactivators analyzed. This induction was particularly strong with the coactivators of the p160 family (7- to 10-fold) and only modest with DRIP205 (2-fold). CITCO did not influence the interaction properties of CAR isoforms SV3, SV4, SV5 and SV6. In contrast, interaction of CAR(SV2) with coactivators was strongly induced. The induction was remarkably strong for the interaction with DRIP205 (100-fold), thereby reaching levels of interaction which were obtained with reference CAR(SV1) in the absence of CITCO.

Besides the very strong expression in liver, mRNA expression of human CAR was reported in heart, muscle, kidney and lung by Northern blot analysis 4. Consequently, we were interested in the expression pattern of CAR splicing variants in different tissues. As transcript SV2 appeared to represent the most frequent splicing variant among the sequenced CAR cDNA clones, we analyzed the tissue expression pattern of transcripts containing an lengthened exon 8. The assay using primers F1/R1 (Table 1) allowed us to distinguish between transcripts with and without the 15 bp insertion in front of exon 8, represented by fragments of 136 bp and 121 bp respectively (Fig. 6A). Transcripts with the 15 bp insertion were detected in fetal liver, kidney, adult liver, lung, trachea and small intestine (Fig. 6A). In contrast CAR transcripts without this insertion could be detected in every tissue analyzed. RT-PCR with primers F3/R3 (Table 1), targeting exons 2/3 which encode parts of the DBD, confirmed the ubiquitous expression of CAR transcripts in human tissues (data not shown).

Fig. 7A shows that CAR expression cannot be detected by Northern blot in intestinal cell lines including subconfluent and confluent Caco-2 TC7. However, if the latter cells were grown for 15 days post confluence, a very strong expression of CAR was observed. As enterocytic differentiation of Caco-2 TC7 cells took place when the cell culture reached confluence 18, we assumed that CAR expression was induced by differentiation. Enterocytic differentiation of post-confluent Caco-2 cells was confirmed by analyzing the expression of the intestinal differentiation marker genes sucrase-isomaltase and alkaline phosphatase. Expression of both genes was strongly induced by post-confluent growth (data not shown). Since we could demonstrate tissue-specific alternative splicing of CAR, it was of interest to see whether alternative splicing may also be associated with the induction of CAR expression by differentiation. RT-PCR with primers F2/R2 clearly demonstrated that confluent, undifferentiated Caco-2 cells predominantly express transcripts with the deletion in exon 9. However, Caco-2 cells grown for 15 days post confluence showed a strong expression of transcripts without and a markedly reduced expression of transcript with the deletion in exon 9 (Fig. 7B). Thus, the strong induction of CAR expression by enterocytic differentiation was also associated with a switch in alternative splicing.

All cell culture media, supplements and fetal calf serum were obtained from Invitrogen (Carlsbad, CA). 6-(4-chlorophenyl)imidazo [2,1-b]thiazole-5-carbaldehyde O-(3,4-dichlorobenzyl)oxime (CITCO) was obtained from Biomol (Plymouth, PA).

The open reading frame of CAR (bases 158–1205 of Genbank accession number NM_005122 / GI 32189358) was amplified by PCR from oligo(dT) primed cDNA samples of an individual human liver and differentiated Caco-2 cells with primers 5'-ATG AAT TCC ACC ATG GCC AGT AGG GAA GAT GAG CTG-3', which introduced an EcoRI site and an optimized Kozak consensus sequence, and 5'-CGT CTA GAT TAG CTG CAG ATC TCC TGG AGC AG-3', which introduced an XbaI site and modified the stop codon to TAA. PCR products of approximately 1000–1100 bp were purified and digested with EcoRI and XbaI. The digested fragment was cloned into appropriately digested pcDNA3 vector (Invitrogen) and the resulting clones were sequenced. Thus we obtained eukaryotic expression plasmids (pcDhCAR) for transcripts SV1, SV2, SV4 and SV6. Part of exons 8–9 (bases 970–1266) was amplified by PCR from oligo(dT)-primed cDNA of a human liver with primers 5'-AAT GAA TTC ACC GAC CTG GAG TTA CCC-3' and 5'-AAT AAG CTT TCC CAC TCC AGT GTA TCC-3'. The oligonucleotides introduced an EcoRI and a HindIII restriction site on the 5'- and 3'-ends of the amplified fragment, respectively. PCR fragments of 200–300 bp were purified, digested with EcoRI and HindIII and subsequently cloned into appropriately digested vector pGEM-7Zf(+) (Promega, Madison, WI). Resulting clones were sequenced.

The eukaryotic expression plasmid for CAR(SV5) was constructed by cloning the 910 bp EcoRI/XmaI insert fragment of pcDhCAR(SV1) and the 130 bp XmaI/HindIII insert fragment of a pGEM exon 8–9 clone (see above) containing the deletion in exon 9 into appropriately digested vector pcDNA3.1(-) (Invitrogen). The internal XmaI site resides at position 1067 of the CAR cDNA. To construct a eukaryotic expression plasmid for transcript SV3, the 15 bp insert lengthening exon 8 was deleted from pcDhCAR(SV6) by use of the QuikChange site-directed mutagenesis kit (Stratagene, La Jolla, CA) according to the recommendations of the manufacturer. The introduction of the deletion and the absence of other undesired mutations were verified by sequencing.

The regions encoding the LBD (amino acids 105–348 of the reference sequence) of CAR protein variants SV1, SV2, SV3, SV4 and SV6 were amplified by PCR with primers 5'-TTA GAA TTC CAA CTG AGT AAG GAG CAA GAA GAG-3' and 5'-TTA TCT AGA GCT GCA GAT CTC CTG GAG C-3', using the respective expression plasmids as templates. The region encoding the LBD of CAR(SV5) was amplified by PCR with the same forward primer and reverse primer 5'-TTA TCT AGA TGG CAG ACA GGC CCT GGA-3' using pcDhCAR(SV5) as a template. The primers introduced EcoRI and XbaI restriction sites on the 5'- and 3'-ends of the amplified fragment, respectively. The EcoRI /XbaI-digested PCR fragments were cloned into appropriately digested vector pVP16 (BD Biosciences Clontech, Palo Alto, CA), thus generating fusion proteins of the VP16 activation domain and the respective CAR LBD. All constructs were verified by sequencing.

The sequence encoding the LBD of human RXRα (amino acids 226–462) was amplified by PCR using 5'-TTA GAA TTC GCC AAC GAG GAC ATG CCG-3', which introduced an EcoRI site and 5'-TTA TCT AGA AGT CAT TTG GTG CGG CGC C-3', which introduced an XbaI site. The resulting PCR fragment was digested with EcoRI and XbaI and cloned into appropriately digested vector pM (BD Biosciences Clontech). The sequences encoding the receptor interaction domains (RID) of human coactivators SRC-1, TIF-2, ACTR and DRIP205 were amplified by PCR from liver cDNA using appropriate primers and cloned into vector pM (BD Biosciences Clontech). The cloned regions encompassed amino acids 583–783 of SRC-1, 583–779 of TIF-2, 582–782 of ACTR and 527–774 of DRIP205. All the resulting plasmids encode fusion proteins of the GAL4-DNA binding domain and the respective regions. The identities of the cloned DNA fragments were verified by sequencing.

The reporter gene construct pGL3-G5 was generated by PCR amplification of the five consensus GAL4 binding sites and adenovirus E1b minimal promoter of pG5CAT (BD Biosciences Clontech) and subsequent cloning into pGL3-Basic (Promega). Enhancer/promoter reporter gene plasmids of CYP2B6 (pB-1.6k/PB/XREM), CYP3A4 (pGL3-CYP3A4(-7830Δ7208-364)) and MDR1 (p-7975/7013/Tk) have been described previously 282930. pB-1.6k/PB/XREM was kindly provided by E. L. LeCluyse (School of Pharmacy, University of North Carolina, Chapel Hill, NC).

COS1 cells were cultivated in Dulbecco's modified Eagle medium buffered with 25 mM HEPES, supplemented with 100 units/ml penicillin and 100 mg/ml streptomycin, 2mM L-glutamine, and 10 % fetal calf serum. The human colon carcinoma cell lines LS174T and LS180 were obtained from the American Type Culture Collection (Manassas, VA). The human colon carcinoma cell line Caco-2 TC7 31 and the human small intestinal cell line BN 32 were kindly provided by U. Meyer (Division of Pharmacology/Neurobiology, Biozentrum of the University of Basel, Switzerland) and G. Pang (Royal Newcastle Hospital, University of Newcastle, Australia), respectively. The human hepatoblastoma cell line HepG2 was obtained from B. Sperker (Institute of Pharmacology, University of Greifswald, Germany). LS174T, LS180, Caco-2 TC7 and BN were cultivated in Dulbecco's modified Eagle medium buffered with 25 mM HEPES, supplemented with 100 units/ml penicillin and 100 mg/ml streptomycin, 1% non-essential amino acids, 1 mM sodium pyruvate, 2 mM L-glutamine and 10% fetal calf serum. HepG2 cells were grown in Minimal Essential Medium, supplemented with 100 units/ml penicillin and 100 mg/ml streptomycin, 2 mM L-glutamine and 10% fetal calf serum. Cells were grown at 37°C and 5% CO₂ in a humidified incubator.

For differentiation experiments, Caco-2 TC7 cells were plated on collagen I coated dishes (BD Biosciences Discovery Labware, Bedford, MA) and grown in standard growth medium until confluence was reached. Then, the medium was replaced by differentiation medium, containing 5% fetal calf serum and culture was continued for 15 days with daily exchange of differentiation medium.

One day before transfection, COS1 cells were plated in 24-well plates (BD Biosciences Discovery Labware) at a density of 3 × 10⁴ cells/well. Transient transfections were performed as described 33. For Two-Hybrid assays, cells were transfected with 110 ng of the reporter gene plasmid pGL3-G5, 10 ng of expression plasmids encoding GAL4-DBD/RXRα-LBD or GAL4-DBD/coactivator-RID fusions, 80 ng of expression plasmids encoding the respective VP16-AD/CAR-LBD fusions and 20 ng of β-galactosidase reference plasmid pCMVβ (BD Biosciences Clontech) per well. In some experiments, cells were subsequently treated with 1 μM CITCO dissolved in Me₂SO or with an equivalent amount (0.1%) of Me₂SO only. Luciferase and β-galactosidase activities were analyzed as described 33. To identify statistical significant differences, one-way analysis of variance with Student-Newman-Keuls post test was performed with the mean values of at least three independent experiments done in triplicates using GraphPad InStat version 3.05 for Windows 95 (GraphPad Software, San Diego, CA).

Electrophoretic mobility shift assays were performed as previously described 30. Human CAR isoforms and RXRα protein were synthesized using the CAR isoform expression plasmids and pCMX-hRXRα (kindly provided by R. Schüle, Klinik für Tumorbiologie, University of Freiburg, Germany), respectively and the TNT T7 Quick Coupled transcription/translation system (Promega). Oligonucleotides for the CYP3A4 DR3-XREM motif were as follows: sense, 5'-GAT CCG CAG AGG GTC AGC AAG TTC ATT CAG A-3'; antisense, 5'-GAT CTC TGA ATG AAC TTG CTG ACC CTC TGC G-3'. Retarded complexes were quantified with the BAS1800 II phosphor-storage scanner (Fuji, Kanagawa, Japan) and AIDA software (Raytest, Straubenhardt, Germany).

For Western blot analysis of CAR isoforms, COS1 cells were co-transfected with the respective expression plasmids and β-galactosidase plasmid pCMVβ (Clontech) using calcium phosphate co-precipitation as described 30. Cells were harvested 24 hours after transfection and cell pellets were resuspended in 10 mM Tris-Cl pH 7.8, 5mM KCl, 2mM MgCl₂. Cell membranes were lysed by addition of Nonidet P-40 to 0.5%, and nuclei were collected by centrifugation. The cytoplasmic proteins in the supernatant were precipitated with acetone. Nuclei and precipitated cytoplasmic proteins were boiled in 62.5 mm Tris-Cl pH 6.8, 10% glycerol, 2 % sodium dodecyl sulfate, 5% 2-mercaptoethanol, 0.02% bromphenol blue and protein samples were resolved on a sodium dodecyl sulfate-10% polyacrylamide protein gel. Western blotting was performed as described 29. CAR proteins were detected with a CAR-specific polyclonal antibody 7, kindly provided by M. Negishi (National Institute of Environmental Health Sciences, Research Triangle Park, NC). Immunoreactive bands were visualized using the chemiluminescence substrate SuperSignal West Dura Extended Duration (Pierce, St Augustin, Germany) and the digital CCD-camera LAS-1000 (Fuji). Results were analyzed with AIDA software (Raytest).

CAR proteins were labeled with [³⁵S] methionine by in vitro synthesis using the TNT T7 Quick Coupled transcription/translation system (Promega) and separated on a sodium dodecyl sulfate-10% polyacrylamide protein gel. The gel was stained with Coomassie Brilliant Blue, incubated in 0.5 M sodium salicylate and dried. The dried gel was exposed to phosphor-storage imaging plates and quantified with the BAS 1800 II phosphor-storage scanner (Fuji) and AIDA software (Raytest).

Total RNA was prepared from human liver samples and small intestine and colon mucosa samples using the RNeasy Mini Kit (Qiagen, Hilden, Germany) as recommended by the manufacturer. The origin of the human liver and intestine samples has been described previously 33. RNA was treated with DNase I to remove contaminating genomic DNA.

5 μg total or polyadenylated RNA was reverse-transcribed with Superscript II reverse transcriptase (Invitrogen) with random hexamer primers according to the standard protocol of the manufacturer.

Expression of CAR transcripts was analyzed by PCR with the primers listed in Table 1. PCR was performed with cDNA corresponding to 500 ng RNA, 1x PCR buffer with 1.5 mM MgCl₂ (Roche Diagnostics, Mannheim, Germany), 200 μM each of dATP, dCTP, dGTP and dTTP, 1 μM each of the forward and reverse primers, and 1.25 U Taq polymerase (Roche Diagnostics). After an initial denaturation step at 94°C for 2 min, cycling conditions comprised 30 cycles of 94°C for 30 sec, 65°C for 30 sec, and 72°C for 30 sec, followed by a final extension step of 72°C for 5 min. PCR products were separated on 10 % polyacrylamide gels and visualized by staining with ethidium bromide.

Preparation of polyadenylated RNA and Northern blot analysis were done as described previously 30. Radioactive probes were synthesized using the DECAprime II Kit (Ambion, Austin, TX). To synthesize a probe specific for human CAR, the cDNA fragment of pcDhCAR(SV1) was used. The origin of the probe specific for human GAPDH has been described 30.