Introduction

Top Summary Introduction Materials & Methods Results Discussion References

TCDD and related aromatic hydrocarbons are important environmental toxicants. Because TCDD is the most potent compound in this chemical class, it has become the prototype for all others. In laboratory animals, TCDD brings about a wide spectrum of toxic or adaptive biochemical and morphological changes, including acute lethality associated with a wasting syndrome; tissue-dependent induction or inhibition of a variety of enzyme activities (the most extensively studied being induction of hepatic CYP1A1); endocrine perturbations; immune suppression coupled with thymic atrophy; lipid peroxidation; alterations in growth factors and cytokines and in their receptors; disturbed synthesis and degradation of heme; and terato- and carcinogenicity (for a review, see Pohjanvirta and Tuomisto, 1994).

The biological effects of TCDD are mediated by an intracellular protein, the AhR (Fernandez-Salquero et al., 1995, 1996; Mimura et al., 1997), which is a member of the bHLH-PAS protein family (for a review, see Okey et al., 1994; Schmidt and Bradfield, 1996). The molecular mechanism of CYP1A1 induction by TCDD has been elucidated in detail. TCDD has the highest affinity of all ligands for the AhR (Safe, 1986). Upon binding TCDD, the AhR sheds its chaperones, two hsp90 molecules, along with other associated proteins and translocates into the nucleus. There it dimerizes with another type of bHLH protein, the ARNT, which is a common partner for several bHLH-PAS proteins (Gradin et al., 1996; Probst et al., 1997). The AhR-ARNT-TCDD complex then binds to DNA at one of several specific sites designated dioxin responsive elements acting as transcriptional enhancers (Whitlock, 1990). This binding induces transcription of the gene CYP1A1 and probably other genes under the regulation of the system by a mechanism not yet fully clarified but which may involve interactions with basal transcription factors (Rowlands et al., 1996).

The bHLH domain is located at the amino terminus of the AhR protein and is responsible for DNA binding and also contributes to protein-protein heterodimerization. The PAS domain participates in dimerization and contains most of the ligand binding domain (Dolwick et al., 1993). Toward the carboxyl terminus, there is a strong trans-activation domain consisting of several subdomains with synergistic action (Jain et al., 1994; Ma et al., 1995).

Regarding acute lethality in guinea pigs, TCDD is the most potent synthetic compound known with an LD₅₀ value of about 1 µg/kg (Schwetz et al., 1973). However, other species are less susceptible; the other extreme, the hamster, tolerates > 1000-fold higher doses (Henck et al., 1981). There are exceptionally wide differences in susceptibility also within species. Two of these, the divergence between inbred C57BL/6 and DBA/2 mice on the one hand and that between inbred L-E (Turku AB) and outbred H/W (Kuopio) rats on the other, have been exhaustively characterized and exploited as biological tools in mechanistic studies.

The sensitivity difference between C57BL/6 and DBA/2 mice is about 10-fold. This is true of acute lethality, CYP1A1 induction, a number of biochemical and histological changes, and teratogenicity (Poland and Knutson, 1982; Chapman and Schiller, 1985). The effects caused by TCDD in the less sensitive DBA/2 strain are qualitatively similar, but require a dose that is an order of magnitude higher to become manifest. In genetic crosses between these strains, sensitivity to AhR ligands is inherited as the dominant trait according to simple Mendelian inheritance for autosomal genes (Gielen et al., 1972). The differential TCDD sensitivity reflects dissimilar AHRs in the two mouse strains. The AhR of the DBA/2 strain has a point mutation (alanine to valine) in its ligand binding domain in codon 375 resulting in markedly reduced binding affinity for TCDD (Ema et al., 1994; Poland et al., 1994); the apparent number of binding sites is also lower than that in C57BL/6 mice (Okey et al., 1989).

The situation is considerably different in the rat model. The LD₅₀ value for L-E rats is only about 10 µg/kg, whereas that for H/W rats is >9600 µg/kg (Pohjanvirta et al., 1993; Unkila et al., 1994). Despite their tremendous resistance to the acute lethality of TCDD, H/W rats are nearly equal to L-E rats in sensitivity to many other biochemical and toxic effects of TCDD, such as induction of CYP1A1, decrease in serum thyroxine and melatonin levels, thymic atrophy and fetotoxicity (Pohjanvirta and Tuomisto, 1994). Also, in contrast to the mouse model, in genetic crosses between these rat strains resistance to TCDD was inherited as a dominant trait, with two or three genes probably involved (Pohjanvirta, 1990). More recent genetic studies have further implied that the AhR gene is the most important of these in terms of sensitivity to the acute lethality of TCDD (Tuomisto et al., 1998).

Recent studies on the AHRs of these two rat strains revealed a notable size dissimilarity. The apparent molecular mass of the AhR, as assessed from Western blots, was 106 kDa in L-E (and also SD rats) but only about 98 kDa in H/W rats (Pohjanvirta et al., 1998). Although the H/W rat AhR showed binding affinity to TCDD similar to that of L-E rat AhR under standard conditions, metal oxyanions (molybdate, tungstate) as well as a high-ionic strength buffer markedly diminished or totally abolished ligand binding of the H/W rat AhR but had little effect on the L-E receptor.

In the present study, we sought to determine the reason for the small size and the physicochemically deviant nature of the H/W rat AhR by molecular cloning and sequencing of the entire coding region of the AhR cDNA. The AhR of the SD rat has previously been cloned (Elferink and Whitlock, 1994; Carver et al., 1994), but for verification of the changes found, representative segments were also cloned and sequenced from L-E and SD rat samples.

	Materials and Methods

Top Summary Introduction Materials & Methods Results Discussion References

Samples. Male rats (female for SD) at the age of 13-14 weeks were used. For validation of the findings, additional male H/W rats that were 5 weeks old were studied. The rats were killed by decapitation. A piece of liver weighing approximately 1 g was rapidly removed by a sterile technique, flash-frozen in liquid nitrogen, and stored at 80° until analysis.

RT-PCR. The liver samples were homogenized while frozen and total RNA was isolated from the homogenate by the phenol-chloroform-guanidine isothiocyanate method (Chomczynski and Sacchi, 1987) using TRIzol reagent (Life Technologies, Eggenstein, Germany). RT-PCR was performed with avian myeloblastosis virus reverse transcriptase (Boehringer-Mannheim, Mannheim, Germany) and either Taq (Pharmacia Biotech, Uppsala, Sweden) or AmpliTaq Gold (Perkin-Elmer-Roche, Branchburg, NJ, USA) DNA polymerase. Random primers were used for RT and the reaction was allowed to proceed at 42° for 30 min. For PCR, the primers were designed on the basis of a published sequence for SD rat AhR (Elferink and Whitlock, 1994). According to this paper, rat AhR cDNA has a 30-NT leader sequence, a 2562-NT coding region (NTs 31-2592) and a total size of 6.6 kb. DNA was amplified with Uno II thermocycler (Biometra, Göttingen, Germany). Thirty-five cycles were always run, but other PCR conditions were optimized for each primer pair. The coding region of the H/W rat AhR cDNA was obtained in five fragments (from 5' to 3' end: 2 or 31 to 839; 810 to 1660; 1561 or 1884 to 2609). The corresponding primers were as follows. Clone 1, 5'-CAGGTGGAGCGGGCA-3' (forward), 5'-AGTGGAGTAGCTATTGCAAACAAAG-3' (reverse); clone 2, 5'-ATGAGCAGCGGCGCCAACATCACCTAT-3' (f), 5'-AGTGGAGTAGCTATTGCAAACAAAG-3' (r); clone 3, 5'-GTTGGCTTTGTTTGCAATAGCTACT-3' (f), 5'-TCCCTAGGTTTCTCATGATGCTATAC-3' (r); clone 4, 5'-CACACTCAGGACGTGAACCTTA-3' (f); 5'-TTCATCCTGGCCTCGAGC-3' (r); clone 5, 5'-GCAGCTTCAGCAGCAGCA-3' (f), 5'-TTCATCCTGGCCTCGAGC-3' (r). A few additional fragments were cloned (e.g., to resolve compression regions).

Genomic DNA. To determine the primary structure of intron 10, liver pieces of about 25 mg were lysed with proteinase K and genomic DNA isolated as 30-50 kDa fragments by spin chromatography (QIAamp Tissue Kit; Qiagen). Intron 10 was then amplified by PCR with the primers 5'-GGTCAGTCCTCAGGCGTACTA-3' (f; 2325-2345) and 5'-GTCCAACCCTCACAGTTCT-3' (r; 2655-2637). The cycling conditions were: 94.8° for 2 min 20 sec; 94.8° for 1 min, 58.7° for 1 min, 72° for 2 min 40 sec × 35; 72° for 12 min.

Cloning and sequencing. The PCR products were cloned by blunt-end cloning into pCR-Script SK(+)Amp plasmids with which XLB-1 supercompetent cells were transformed (Stratagene, La Jolla, CA). The plasmids were purified by Wizard Plus (Promega, Madison, WI) or Quantum Prep (Bio-Rad, Hercules, CA) miniprep systems. The inserts were sequenced with an A:L:F-automated DNA sequencer (Pharmacia Biotech) using either AutoRead sequencing kit (Pharmacia Biotech) or Thermo Sequenase fluorescent-labeled primer cycle sequencing kit (Amersham, Cleveland, OH).

Cloning strategy. Each clone was first sequenced from an H/W rat. If any differences from the published SD rat receptor base order were detected (or if there were compression areas), the reciprocal strand of the DNA double helix was analyzed on the same site or, if the clone was too large for this, another fragment over the same region was amplified with slightly different primers. If the alteration was confirmed, this region was examined in another H/W rat. If an identical change was found, its strain specificity was assessed by cloning and sequencing this region from L-E and SD rats (intron 10 was only cloned from L-E and H/W rats). The entire 3' end (NTs 1561 to 2609) of the AhR protein coding region was cloned and sequenced from all three strains. As an extra measure, a large portion of the 5' end (NTs 127-839) was also cloned from L-E and SD rats. Thus, the sections analyzed for each strain were as follows: H/W, NTs 2-2609 and intron 10; L-E, NTs 127-2609 and intron 10; SD, NTs 127-839, 1561-2609.

Southern blotting. After separation in a 1% agarose gel, PCR-amplified DNA stretches were denatured by submerging the gel in 0.5N NaOH, 1.5 M NaCl solution for 30 min followed by neutralization with 0.5 M Tris·HCl, 3 M NaCl, pH 7.5, for another 30 min. The PCR products were then transferred from the gel onto a positively charged nylon membrane (Hybond+; Amersham) by capillary diffusion overnight. DNA was immobilized by baking the membrane at 120° for 30 min. Prehybridization and hybridization were accomplished in bottles in a rotisserie-equipped hybridization oven (Hybaid, Teddington, UK) at 68°. The probe used was the antisense strand for the shorter insertion of H/W rat AhR (see Results) labeled with digoxigenin at its 5' end. Detection was performed with the colorimetric detection reagents 2,2'-di-p-nitrophenyl-5,5'-diphenyl-3,3'-[3,3'-dimethoxy-4,4-diphenylene]ditetrazolium chloride/-bromo-4-chloro-3-inodolyl phosphate (NBT/BCIP) according to manufacture's instructions (Boehringer-Mannheim).

	Results

Top Summary Introduction Materials & Methods Results Discussion References

The 5' end of the H/W rat AhR, comprising the bHLH and PAS domains (NTs 109-276 and 358-1185, respectively), proved to be identical to that of the previously published sequence for SD rats (Elferink and Whitlock, 1994; Carver et al., 1994). This was also true of the fragment (NTs 127-839) cloned from L-E and SD rats.

A transition mutation was recorded in H/W rats in the variable region flanking downstream the PAS domain. The second base of codon 497 (NT 1520) was altered from T to C, resulting in an amino acid change from valine to alanine at the protein level. This point mutation was strain specific, because it did not occur in L-E rats that had the same sequence published previously for SD rats.

A more conspicuous change was found toward the 3' end of the coding region of the cDNA. Although the reverse-transcribed RNA from L-E and SD rat livers always provided the expected single band in the subsequent PCR amplification, three distinct bands were constantly detected in the case of H/W rats (Fig. 1). One of the bands was about 100 bases smaller than the expected size. The other two were some 30 and 150 bases larger than expected. These cDNA stretches were isolated, purified, and reamplified by PCR. They were then individually cloned and sequenced.

View larger version (105K): [in this window] [in a new window]   Fig. 1.   PCR amplification products from the 3' end of the rat AH receptor coding region. The forward and reverse primers recognized bases 1884-1901 and 2609-2592, respectively, and thus the expected size of the PCR product was 726 nucleotides. Lane 1, 1 kb DNA molecular mass marker (MBI Fermentas, Vilnius, Lithuania); lanes 2-4, H/W rats; lane 5, SD rat; lane 6, Product 5 diluted 1:5 (to show the exact size). The rat samples used in PCR were RT products obtained with random primers from 1 µg total RNA (2.5 µl/50 µl PCR reaction volume).

The bottom band turned out to contain a deletion of 129 bases (NTs 2326-2454). The ends of this deletion matched codon boundaries; therefore, it would not lead to a frame-shift in the translated protein, but only a loss of 43 amino acids (AAs 766-808). The middle band was the expected sequence with an insertion of 29 bases between NTs 2454 and 2455. It contained a termination codon (UAA) in its sequence and would give rise to an AhR protein with the last 45 carboxyl-terminal amino acids missing (AAs 809-853) and seven new ones added to its carboxyl terminus, for a total loss of 38 amino acids. The top band represented the expected DNA stretch with an insertion of 134 bases between the same NTs 2454 and 2455. At its 5' end, it encompassed the entire sequence of the smaller insertion. Because of the stop codon included, this mRNA would translate into the same protein as the 29-base insertion. We will call the three corresponding full-length cDNA structures in H/W rats H/Wdv, H/Wsiv, and H/Wliv for deletion variant, shorter insertion variant and longer insertion variant, respectively. The departures of the H/W rat AhR primary structure from that of SD and L-E rats are summarized in Fig. 2.

View larger version (22K): [in this window] [in a new window]   Fig. 2.   Schematic summary of the changes recorded in H/W rat AhR cDNA versus "wild type" AhR in L-E and SD rats. ORF, open reading frame.

To further confirm the findings, a digoxigenin-labeled probe of the antisense strand for the smaller insertion was produced. Southern blots revealed that it avidly recognized two bands in cDNA amplified from the 3' end of H/W rat AhR coding region but no bands in the case of L-E or SD rats (Fig. 3; only SD shown). Of the isolated cDNA stretches from H/W rats, the probe strongly labeled the two largest (data not shown). When the digoxigenin-labeled DNA fragment was exploited as the reverse primer in RT-PCR [with the 18-mer 5'-GCAGCTTCAGCAGCAGCA-3' (NTs 1884-1901) as the forward primer), a strong signal was obtained from poly A⁺ mRNA in the case of H/W rats but a very faint one (possibly originating from genomic DNA traces or nuclear RNA) in the case of L-E rats. From genomic DNA, L-E rat samples yielded a signal similar to that in H/W rats (Fig. 4).

View larger version (104K): [in this window] [in a new window]   Fig. 3.   Southern blot of the gel shown in Fig. 1. The probe used was digoxigenin-labeled antisense strand for the shorter insertion. The probe recognizes only H/W samples.

View larger version (91K): [in this window] [in a new window]   Fig. 4.   PCR amplification products from reverse-transcribed poly A+ mRNA (lanes 2-3) or from genomic DNA (lanes 4-5). The reverse primer was digoxigenin-labeled antisense strand for the shorter insertion fragment from H/W rat AhR cDNA (the first 29 NTs of rat AhR intron 10 with a G-to-A mutation at the 5' end) and the forward primer was the 18-mer 1884-1901 from the rat AhR cDNA. Lane 1, 1 kb marker (MBI Fermentas); lanes 2 and 4, samples from the same H/W rat; lanes 3 and 5, samples from the same L-E rat. Note that the amplified segment contains the possible supernumerary exon/intron junction at 2325-2326.

All three changes were thus centered at NT 2454. The genomic organization of the AhR gene has so far only been elucidated in the mouse (Schmidt et al., 1993). Comparison of the mouse and rat AhR sequences reveals that the position 2454 represents the 3' end of exon 10 in the rat. This suggested that the three products in H/W rats might be attributable to alternative mRNA splicing. This, in turn, aroused the idea that the 129-base deletion (H/Wdv) could be a reflection of exon skipping. Exon 10 is a long exon in the mouse consisting of 1264 NTs; it could be split into smaller parts in the rat. To test this hypothesis, genomic DNA was isolated from H/W and L/E rat livers and primers designed to amplify the possible intron between NTs 2325 and 2326. However, all the PCR reactions yielded products whose sizes exactly matched the expected product sizes from cDNA (Fig. 4), thereby disproving the hypothesis.

To find the source of the three products from the 3' end of AhR cDNA in H/W rats, primers were then devised on both sides of NT 2454. PCR amplified a product of about 3.2 kb from both H/W and L-E rat genomic DNA. When these products were subjected to Southern blotting, the 29-base probe (see above) recognized both (data not shown). This prompted us to clone and sequence these introns.

The two insertion fragments in H/W rat AhR cDNA proved to emanate from intron 10 (Fig. 5). Moreover, the intron 10 sequence was found to be highly homologous between L-E and H/W rats. In the 5' end, there was only one difference within the first 150 NTs: the first base, the invariant G, was substituted by A in H/W rats. At the 3' end, the first divergence was at NT 69. The total number of nonmatching bases was 6/485 and 6/439 for the 5' and 3' ends, respectively. Both ends of intron 10 also shared high homology between rat and mouse. When a stretch equal to the size of the larger insertion sequence (134 NTs) was analyzed in the 5' end, it proved to be 75% identical over its entire length with all the first 12 NTs being the same in both species (Fig. 6a). At the 3' end, the first three NTs were identical and there was a 56-NT segment of remarkable homology (84%) a little farther away (Fig. 6b).

View larger version (23K): [in this window] [in a new window]   Fig. 5.   Primary structure of the 3' end of exon 10 and 5' end of intron 10 in the rat AhR gene. The deviations occurring in H/W rats are depicted.

View larger version (33K): [in this window] [in a new window]   Fig. 6.   Alignment for homology between rat and mouse intron 10 sequences at the 5' (a) and the 3' (b) ends. Lines, identical bases.

Examination of the exon/intron 10 junctions revealed the reason for the three alternative cDNAs in the H/W rat. The wild type junction occurring in L-E and SD rats has the invariant GT dinucleotide at the 5' end of intron 10, and the overall correspondence with the vertebrate consensus sequence (Padgett et al., 1986) is high (Fig. 7). The transition mutation in H/W rats destroys the critical recognition site, which leads to a search of cryptic splice sites. The first site upstream meeting the requirements for an exon/intron junction is at position 129 (between NTs 2325 and 2326), and usage of this site results in H/Wdv. The first such site downstream occurs at position +29 giving rise to H/Wsiv. This site is only weakly matching with the consensus on the "exon" side (consensus: A/C-A/C-A-G; +29: C-C-T-T), which may account for the concurrent usage of the next possible splice site at position +134, which has greater similarity to the consensus sequence (T-C-A-G). H/Wliv arises from splicing at this position.

View larger version (25K): [in this window] [in a new window]   Fig. 7.   Splice site structures at the exon/intron 10 junction in comparison with the vertebrate consensus sequence (Padgett et al., 1986). The relative frequencies of occurrence for NTs at each position in the consensus sequence are shown. Bold, invariant GT dinucleotide; lower-case letters, NTs deviating from the consensus sequence.

The noncoding regions of the AhR cDNA were similar in size in all the three rat strains studied, because Northern blots yielded in each case a product of about 6.6 kb (data not shown). However, it should be noted that the resolution power of this method is very limited, and the different products in H/W rats could not be distinguished.

The two previous publications on the amino acid sequence of rat AhR differ from one another at one position, 794, which was reported to be histidine by Elferink and Whitlock (1994) for SD rats but glutamine by Carver et al. (1994) for Fisher 344 rats. At the corresponding cDNA NT position, 2412, we obtained G for all three strains of rat studied, including SD, instead of T reported by Elferink and Whitlock (1994). This G would convert the codon to that of glutamine. Therefore, the difference is probably attributable to an error in amplification or sequencing in the Elferink and Whitlock (1994) paper.

	Discussion

Top Summary Introduction Materials & Methods Results Discussion References

The H/W strain stands out from all other rat strains for its exceptional TCDD resistance. Previous studies have shown that this resistance is not due to altered overall kinetics of TCDD (Pohjanvirta et al., 1990), which may contribute to the 5-fold sensitivity variation among substrains of Long-Evans rats (Viluksela et al., 1996). Involvement of the AhR in the resistance was suggested earlier by two lines of evidence. Firstly, both strains were equally susceptible to the acute lethality of perfluorodecanoic acid (Unkila et al., 1992), which causes a wasting syndrome similar to that of TCDD but does not act through the AhR. Secondly, the fold-difference in LD₅₀ values between the strains narrows in a step-by-step manner, with an increase in the number of chlorine atoms in the dioxin molecule, as follows: >1000, 320, 20, 10 for the most potent congeners of tetra-, penta-, hexa- and heptachlorodibenzo-p-dioxins, respectively (Pohjanvirta et al., 1995).

In support of this line of reasoning, the AHRs of the two strains were found to have dissimilar molecular masses by immunoblotting (Pohjanvirta et al., 1998). Very recent genetic crossing and molecular biological studies imply that the smaller size (~98 kDa) of H/W rat AhR [compared with that of L-E rat AhR (~106 kDa)] stems from the altered H/W rat AhR cDNA structure described here and that rats with the deviant AhR are about 50 times more resistant to TCDD lethality than those with the wild-type receptor (Tuomisto et al., 1998; Pohjanvirta R, unpublished observations). However, the unresponsiveness of H/W rats to TCDD is peculiarly endpoint-dependent, and they fully exhibit many classic AhR-mediated effects, such as CYP1A1 induction (Pohjanvirta et al., 1988). Recent studies further demonstrate that there is no discernible difference in the affinity of ligand binding between the AHRs of L-E and H/W rats, although receptor density is somewhat (2-3 times) higher in L-E rats; the ARNT proteins of the two strains are also similar in size (Pohjanvirta et al., 1998).

Collectively, the data gathered on H/W rats imply that their AhR is somehow deviant, but the domains responsible for ligand and DNA binding as well as heterodimerization are probably intact. In full accordance with this hypothesis, we now found that the entire 5' end of the H/W rat AhR cDNA encompassing these conserved domains is identical to that in other strains. This clearly distinguishes the rat model from its mouse counterpart at the molecular level, because the higher resistance to TCDD of DBA/2 mice compared with C57BL/6 mice results from reduced binding of TCDD to DBA/2 AHRs due to a single amino acid change (alanine to valine) within the ligand binding domain (Ema et al., 1994; Poland et al., 1994).

The segment that showed a striking departure from the normal rat AhR structure in H/W rats was the 3' end of the cDNA coding region. A transition mutation (G to A) in the first invariant NT of intron 10 abolishes the default exon/intron junction leading to the usage of three nearest cryptic splice sites. This in turn gives rise to a deletion and two insertions, the smaller insertion being the 5' end of the larger. At the protein level, the two insertions would yield the same product, an AhR variant that lacks the last 38 amino acids from its carboxyl terminus and has an altered sequence for seven amino acids in this end. The deletion would translate into a protein with 43 amino acids missing at a site 45 amino acids away from the carboxyl terminus. The calculated molecular masses of these deficiencies are 5445 (H/Wdv) and 4572 Da (H/Wsiv and H/Wliv). They are slightly smaller than the figure obtained for the size difference from immunoblots (~8000 Da), but the match is seldom perfect. For example, the calculated molecular mass of the SD rat AhR is 96,231 Da, but immunoblots consistently give a value of 106 kDa (Elferink and Whitlock, 1994; Pohjanvirta et al., 1998), possibly reflecting aberrant mobility during sodium dodecyl sulfate-polyacrylamide gel electrophoresis. The two predicted protein products in H/W rats are so close to one another in size that at the moment it is unknown whether both of them are really synthesized in vivo. Verifying this will require raising specific antibodies against the missing parts.

The insertion-based missing segment is rich in leucine (15.6%), alanine (13.3%) serine (13.3%) and proline (11.1%). The skipped part stemming from deletion is rich in glutamine (16.3%), alanine, proline and serine (all three 11.6%). They both belong to the trans-activation domain of the AhR (Jain et al., 1994; Whitelaw et al., 1994; Ma et al., 1995), which may be reflected in the high glutamine content in the latter defect. Studies on the mouse AhR have revealed that the carboxyl-terminal end of the AhR actually consists of several subdomains with synergistic coeffects (Ma et al., 1995). It is conceivable that the contribution of each subdomain to regulation of gene activity varies from gene to gene, which might account for the selective responsiveness of H/W rats to TCDD.

Another mechanism for the unique pattern of responsiveness in H/W rats could be that the gene(s) responsible for acute lethality of TCDD are predominantly controlled by the AhR partner of the AhR/ARNT complex, whereas for some other dioxin-regulated genes (e.g. CYP1A1) the converse is true. In support of this hypothesis, by using AhR and ARNT chimeras, Whitelaw et al. (1994) showed that relative trans-activational potencies of the two partners varied with cell type and promoter architecture.

A third conceivable mechanism is based on the recent discoveries that the unliganded AhR is associated with the protein kinase c-Src, which becomes activated upon ligand binding (Enan and Matsumura, 1996) and that c-Src deficient mice are resistant to many endpoints of TCDD toxicity including acute lethality (Matsumura et al., 1997). c-Src is usually bound to hsp90 (Pratt, 1997), which makes contact with the AhR at two distinct sites within the bHLH and PAS domains (Perdew and Bradfield, 1996). The metal oxyanions molybdate and tungstate stabilize AhR ligand binding (Manchester et al., 1987; Okey et al., 1989), probably by the same mechanism as in the case of steroid receptors, strengthening the association between hsp90 and AhR (Pratt, 1997). We recently detected, however, that in H/W rats, these metal oxyanions surprisingly labilized the AhR (Pohjanvirta et al., 1998), which suggests that the association between hsp90 and AhR may be altered in these rats. If this holds true, the activation of c-Src may also not take place in a normal manner in H/W rats. In this context, it is intriguing to note that c-src knockout mice, in spite of their resistance to the acute lethality of TCDD, are fully responsive to CYP1A1 induction (Matsumura et al., 1997), as are H/W rats.

The biological significance of the exon 10 mutation is uncertain. The transition mutation at the second base of the codon causes an amino acid shift from valine to alanine at position 497. Because both of them are aliphatic, nonpolar, hydrophobic amino acids and the mutation occurs in an hypervariable region, it is not likely that this change would play a major role in the resistance of H/W rats to TCDD. In fact, the mouse AhR has alanine at the corresponding position.

The high degree of conservation of intron 10 sequences between rat and mouse was a surprising finding. Usually, the only sites in introns showing evolutionary conservation are the splicing junctions (especially the dinucleotides GT... AG) and the branch site located near the 3' end (Lewin, 1994). For example, Hughes and Yeager (1997) studied 42 pairs of orthologous genes from rats and mice and found that the evolutionary rate in introns reflected the mutation rate. In light of this evidence, the 5' end homology of intron 10 is quite striking. It would be interesting to know whether this apparent conservation applies only to intron 10 or whether it is a general phenomenon in the AhR.

The example provided here by H/W rats demonstrates that intron sequences may have functional significance and should not be overlooked in genetic analyses. Krawczak et al. (1992) found that point mutations within splice sites represent some 15% of known point mutations causing human genetic disease. Mutations in the 5' splice site were much more frequent than in the 3' junction. Overall, 81% of all point mutations within the 5' splice site involved substitution of the intron first G residue, and in 64% of these cases, the substituting base was A, as in H/W rats.

In conclusion, the most TCDD-resistant rat strain, H/W, harbors two point mutations in the AhR gene that manifest as alterations in the AhR protein structure. One of them is in exon 10 and results in a change of a single amino acid within a variable region of the AhR. The other mutation is in the first invariant NT at the 5' end of intron 10 and leads to usage of cryptic splice sites to form 3 mRNA species but resulting in only two possible proteins. The variant H/Wdv uses a cryptic splice site between NTs 2325 and 2326 and results in an in-frame deletion at NT 2326 giving rise to a 43-amino acid deletion in exon 10 of the AhR protein. The variant H/Wsiv mRNA results from the use of a cryptic splice site at +29 NTs in intron 10 causing a 29-bp insertion that includes a termination codon (UAA), giving rise to an AhR protein with a total deletion of 38 amino acids. The variant H/Wliv mRNA results from the use of a cryptic splice site at +134 NTs in intron 10; since this includes the termination codon found in H/Wsiv, this mRNA translates into the same protein as for H/Wsiv mRNA. Thus there are two possible H/W AhR proteins deleted of either 43 or 38 amino acids (with altered sequence for the last seven amino acids in the latter case) near the end of the carboxyl terminus of the trans-activation domain. The functional consequences of these changes remain to be characterized.