Introduction

Top Abstract Introduction Materials & Methods Results Discussion References

Maintaining control of hepatic drug-metabolizing enzyme gene expression is essential to the multifaceted process of xenobiotic detoxication. In rat liver, members of the SULT1¹² gene subfamily catalyze the sulfation (or o-sulfonation) of a number of important endogenous and exogenous substrates, ranging from phenols to biogenic amines (Duffel et al., 1991). The SULT2 gene subfamily catalyzes the sulfation of aliphatic alcohols, hydroxysteroids, bile acids, and other conceivably toxic substrates (Jakoby et al., 1980). In metabolically active tissues such as the liver, enzymatic sulfation of xenobiotics may culminate in the excretion of more polar end-products (detoxication). Alternatively, unbridled generation of highly reactive sulfate esters in toxicant target tissues may flood the system with electrophilic species that readily form covalent adducts with critical cellular nucleophiles, such as DNA, RNA, and protein (toxicant bioactivation). In rat liver, the ability of the SULT1 and SULT2 enzymes to metabolize xenobiotics has been suggested to play a key role in hepatocarcinogenesis. For example, SULT2 enzymes bioactivate hydroxymethyl polycyclic aromatic hydrocarbon procarcinogens, such as 5-hydroxymethylchrysene and 7,12-dihydroxymethylbenz[a]anthracene, to reactive electrophilic intermediates (Okuda et al., 1989; Ogura et al., 1990b), and SULT1-mediated bioactivation is thought to amplify the carcinogenicity of N-hydroxy-2-acetylaminofluorene (Yerokun et al., 1992; Meerman et al., 1981).

Members of the SULT1 and SULT2 gene subfamilies represent the principal SULT enzymes that are essential to rat hepatic xenobiotic metabolism. There are at least four SULT1 enzyme isoforms and three individual SULT2 enzyme isoforms present in rat liver (Sakakibara et al., 1995; Yamazoe et al., 1994; Liu et al., 1996). Historically, the SULT1 and SULT2 enzymes have not been classified as "xenobiotic-inducible" enzymes. However, we previously demonstrated that SULT gene expression is affected by aryl hydrocarbon receptor agonist, P450-inducing agents. For example, in vivo treatment with the CYP1A1 inducer 3-methylcholanthrene suppressed rat hepatic SULT2 (HST-a) mRNA expression in a dose-dependent manner (Runge-Morris and Wilusz, 1994). Similarly, we and others have shown that in vivo treatment with CYP3A-inducing doses of xenobiotic steroidal chemicals, such as dexamethasone or pregnenolone-16 -carbonitrile, increases rat hepatic SULT1 and SULT2 mRNA expression (Runge-Morris et al., 1996; Liu and Klaassen, 1996a).

The effects on SULT gene expression of another major class of xenobiotic P450-inducing agents, namely PB and "PB-like" inducers, have not been systematically established. PB and the PB-like agents, such as certain organochlorine compounds, polychlorinated biphenyls, and imidazole antimycotic agents, induce CYP2B1/2 gene expression and cause pleiotropic effects on hepatic gene expression that include increases in P450, glutathione-S-transferase, epoxide hydrolase, and UDP-glucuronosyltransferase enzyme expression (Lubet et al., 1992). In addition, PB may produce divergent effects on closely related drug-metabolizing enzyme isoforms. Preliminary data in our laboratory suggested that in vivo treatment of rats with the archetypical CYP2B1/2 inducer PB produces alterations in hepatic SULT1 and SULT2 gene expression (Runge-Morris and Vento, 1994). In this report, the effects of PB on SULT1 and SULT2 gene subfamily expression and on SULT1/2 isoform-specific mRNA expression in rat liver were characterized.

	Materials and Methods

Top Abstract Introduction Materials & Methods Results Discussion References

Materials. PB and molecular biology-grade reagents were purchased from Sigma Chemical Co. (St. Louis, MO). Trizol reagent and Klenow fragment were obtained from GIBCO BRL (Grand Island, NY). Terminal deoxynucleotidyl transferase was purchased from Promega Corp. (Madison, WI). SDS-polyacrylamide gel electrophoresis reagents were obtained from Bio-Rad Laboratories (Hercules, CA). ECL Western blotting kits and [-³²P]dATP (>3000 Ci/mmol) were purchased from Amersham (Arlington Heights, IL). Nylon hybridization membranes (Genescreen Plus) and [-³²P]dATP (>6000 Ci/mmol) were purchased from Du Pont NEN Co. (Boston, MA).

Animals. Adult male Sprague-Dawley rats (age, ~55 days; 264-280 g) were purchased from Harlan, Inc. (Indianapolis, IN). Rats were housed in steel mesh cages and were supplied with Purina Rodent Chow (Purina, Indianapolis, IN) and water ad libitum before the initiation of experimental protocols.

In Vivo Drug Treatment. Based on preliminary dose-optimization studies, rats were treated for 3 consecutive days with ip injections of control 0.9% saline solution or PB (35, 80, or 100 mg/kg in water). The doses of PB used in this study did not produce signs of clinical toxicity in treated animals. After the final drug treatment, rats were fasted for 24 hr and then killed with lethal pentobarbital injections (120 mg/kg, ip). The livers were then perfused in situ with ice-cold 0.9% saline solution and were powdered under liquid nitrogen. To permit statistical analysis of the data, three rats were included in each experimental group (N = 3).

Origin of cDNA and Oligonucleotide Probes. Rat SULT1 (ASTIV) and SULT2 (HST-a) cDNA probes were generated using RT-PCR, as described previously (Runge-Morris et al., 1996; Runge-Morris and Vento, 1995). The 799-base pair SULT1 cDNA probe was prepared by RT-PCR using primers corresponding to nucleotides 97-112 and 881-896 of the previously published rat hepatic SULT1 cDNA sequence (GenBank accession number X52883) (Yerokun et al., 1992; Ozawa et al., 1990). The 518-base pair SULT2 cDNA probe was prepared by RT-PCR using oligonucleotide primers corresponding to nucleotides 306-321 and 809-824 of the ST-20 SULT2 cDNA sequence (GenBank accession number M31363) (Ogura et al., 1989). The SULT-Dopa/tyrosine oligonucleotide probe used in these studies was selected using Oligo PrimerAnalysis software (National Biosciences Inc., Plymouth, MN) and published cDNA sequence information for rat hepatic SULT-dopa/tyrosine (GenBank accession number U38419) (Sakakibara et al., 1995) (table 1). Oligonucleotide probes specific for the SULT-Dopa/tyrosine isoform and for SULT1 isoforms (SULT1A1, SULT1C1, and SULT1E2) and SULT2 isoforms (SULT20/21, SULT40/41, and SULT60) previously described and characterized by Liu et al. (1996) were synthesized by National Biosciences Inc. Specific sequence and hybridization information on the isoform-specific oligonucleotide probes used in these studies is provided in table 1.

Northern Blot Analyses. Total RNA was prepared from 1-g samples of powdered liver as previously described (Puissant and Houdebine, 1990; Runge-Morris et al., 1996; Runge-Morris and Vento, 1995). Poly(A)⁺ RNA was prepared from 1-mg samples of total RNA using the Qiagen digotex mRNA Midi kit (Chatsworth, CA), in accordance with the manufacturer's instructions, and poly(A)⁺ RNA (2 µg/lane) was analyzed on Northern blots hybridized with isoform-specific SULT oligonucleotide probes. Briefly, Northern blots were prehybridized overnight in a prehybridization/hybridization solution containing 5× Denhardt's solution (0.1% Ficoll, 0.1% polyvinylpyrrolidone, 0.1% bovine serum albumin), 6× SSC (1× SSC contains 150 mM sodium chloride and 15 mM sodium citrate), 1% SDS, 200 µg/ml sheared, hydrolyzed, salmon sperm DNA, and 200 µg/ml polyadenylic acid and were hybridized overnight at the temperature of dissociation minus 4°C-21°C, in prehybridization/hybridization solution containing 20 × 10⁶ cpm/ml levels of individual SULT1 or SULT2 oligonucleotide probes labeled using terminal deoxynucleotidyl transferase (Collins and Hunsaker, 1985). After hybridization, filters hybridized with SULT1E2, SULT1C1, SULT-Dopa/tyrosine, or SULT2 oligonucleotide probes were washed twice at hybridization temperature for 1 hr with 5× SSC/0.5% SDS and once for 1 hr with 5× SSC/0.1% SDS and were exposed to autoradiographic film. Filters hybridized with the SULT1A1 oligonucleotide probe were washed as described above but also received a more stringent wash with 2× SSC/0.1% SDS at the temperature of dissociation minus 9°C for 1 hr. Alternatively, filters were hybridized with labeled SULT1 or SULT2 cDNA probes, washed, and exposed to autoradiographic film as previously described (Runge-Morris et al., 1996). To normalize for RNA loading and transfer, blots were stripped of labeled SULT probes as previously described (Runge-Morris et al., 1996) and were rehybridized with P-labeled -actin cDNA probe (prepared using random-prime labeling).

Western Blot Analyses of Immunoreactive SULT1 and SULT2 Protein Levels in Rat Liver Cytosol. Polyclonal antiserum to rat SULT1 protein was prepared and characterized as described previously (Hirshey et al., 1992; Hirshey and Falany, 1990). Polyclonal antiserum to rat SULT2 (formerly called HST-a) protein was prepared as described previously (Runge-Morris and Wilusz, 1994). Rat hepatic cytosol was prepared from the livers of rats treated in vivo with vehicle or PB, and protein concentrations in cytosol samples were determined according to established methods (Smith et al., 1985). Hepatic cytosol samples (3.0 µg/lane for anti-SULT1 antiserum blots, 7.5 µg/lane for anti-SULT2 antiserum blots) were fractionated by SDS-polyacrylamide gel electrophoresis (Laemmli, 1970), and immunoreactive SULT1 and SULT2 protein levels were determined using ECL Western blot analysis, as described previously (Runge-Morris and Wilusz, 1994; Runge-Morris and Vento, 1995).

Statistical Analysis of the Data. Northern blot autoradiographs and Western blot chemiluminescence films were analyzed by scanning laser densitometry using the ImageQuant software package (Molecular Dynamics, Sunnyvale, CA), as described previously (Runge-Morris et al., 1996). Statistical analysis of Northern blot data was performed using one-way ANOVA, followed by Dunnett's test to compare treatment effects with control. The dose-response data were also analyzed using linear regression, to assess whether the slopes of the best-fit lines were significantly different from 0. Western blot data were analyzed using unpaired t tests. All statistical tests were performed using the Prism software package (GraphPad, San Diego, CA).

	Results

Top Abstract Introduction Materials & Methods Results Discussion References

The effects of in vivo treatment with PB on the mRNA and protein expression of the rat hepatic SULT1 and SULT2 gene subfamilies and on the mRNA expression of individual SULT1 and SULT isoforms were investigated. Preliminary studies in our laboratory suggested that the PB-mediated effects on SULT1 and SULT2 gene expression might be subtle. However, because of the potential for Northern blotting with SULT cDNA probes to mask more demonstrable PB-mediated effects on single SULT isoforms, we determined to assess the effects of PB treatment on the mRNA expression of specific SULT1 and SULT2 isoforms. Gender-dependent differences in rat hepatic SULT1 and SULT2 gene expression (Song et al., 1990; Runge-Morris and Wilusz, 1991; Runge-Morris, 1994; Singer and Sylvester, 1976) and regulation (Ueda et al., 1997; Liu and Klaassen, 1996a,b,c) have been described. The present study focused on characterizing PB-mediated effects on hepatic SULT1 and SULT2 expression in male rats because we found detectable mRNA levels for individual SULT enzyme isoforms in the livers of mature male rats and because our most recent mechanistic studies on the regulation of hepatic SULT1 and SULT2 gene expression by glucocorticoid-class P450 inducers were performed with mature male rats (Runge-Morris et al., 1996).

Northern blots of poly(A)⁺ RNA isolated from the livers of rats treated with PB or saline vehicle were hybridized with a SULT1 cDNA probe that was originally designed to maximize the detection of SULT1A1 and related SULT1 isoforms, as described in Materials and Methods. This experiment revealed that the lowest test dose of PB (35 mg/kg) produced a moderate, but statistically insignificant (by ANOVA and Dunnett's test), increase in rat hepatic SULT1 mRNA levels, to ~138% of vehicle-treated control levels (fig. 1). Higher doses of PB failed to produce substantial alterations in SULT1 mRNA levels, in comparison with vehicle-treated control (fig. 1).

View larger version (34K): [in this window] [in a new window]   Fig. 1.   Effects of in vivo treatments with PB on SULT1 and SULT2 mRNA levels in male rat liver. Adult male Sprague-Dawley rats (three rats/treatment group) were treated for 3 consecutive days with saline vehicle (VC) or PB (35, 80, or 100 mg/kg, ip) and were sacrificed 24 hr after the last treatment. Poly(A)+ RNA was prepared from a sample of each individual liver, and aliquots (2 µg/lane) were analyzed by Northern blot hybridization, using a cDNA probe to SULT1 or SULT2. Hybridizable RNA bands were detected by autoradiography, and band intensities were estimated by scanning laser densitometry. After densitometric analysis, SULT probes were washed from the blots, and the blots were rehybridized with a -actin cDNA probe. A, autoradiographs of Northern blots hybridized with SULT1 or SULT2 cDNA. A representative autoradiograph of a blot rehybridized with -actin cDNA is also shown. B, graphical representations of SULT1 and SULT2 mRNA band intensities. The SULT1 and SULT2 band intensities were normalized to the respective -actin band intensities and are presented as percentages of control (mean ± SD, N = 3). * The slope of the best-fit line for the dose-response data is significantly different from 0 (p < 0.05).

To complement statistical analysis of the data by ANOVA and Dunnett's test, linear regression analyses of the dose-response data were performed, as a simple method to determine the degree to which PB treatments produced dose-dependent increases or decreases in SULT mRNA expression. If linear regression analysis supported a dose-dependent change in SULT mRNA expression in response to PB treatment, the slope of the linear regression best-fit line was significantly different from 0 (indicated as * in figs.). Treatment for 3 consecutive days with increasing doses of PB (35, 80, or 100 mg/kg) tended to suppress rat hepatic SULT2 mRNA expression in a relatively dose-dependent manner, to levels that were ~88, ~62, and ~58% of control levels, respectively (fig. 1), and the slope of the best-fit line approximating this dose-response relationship was significantly less than 0. As expected, PB-treated rats displayed significant increases in the amounts of hepatic CYP2B1/2 mRNA, with maximal CYP2B1/2 induction occurring at a PB dose of 80 mg/kg and being sustained at the 100 mg/kg dose (data not shown).

The effects of PB on immunoreactive SULT1 and SULT2 protein levels were also evaluated, and results were compared with PB-mediated changes in SULT1 and SULT2 mRNA levels. Western blots were performed with rat hepatic cytosolic protein isolated from rats treated with PB or saline vehicle. In the absence of available isoform-specific anti-SULT antibodies, class-specific anti-SULT1 and -SULT2 antisera were used in ECL Western blot analyses. The polyclonal anti-SULT1 antiserum used in these studies was designed to detect rat hepatic "minoxidil SULT," as described previously (Hirshey et al., 1992; Hirshey and Falany, 1990), and is expected to cross-react with SULT1A1 subunit protein and related isoform subunits. The anti-SULT2 antiserum used in these studies was prepared as described previously (Runge-Morris et al., 1996; Runge-Morris and Wilusz, 1994). Given the strong structural sequence similarities among SULT2 subunit proteins, it is likely that our polyclonal anti-SULT2 antiserum detects all of the known SULT2 isoforms, including SULT20/21, SULT40/41, and SULT60 (Watabe et al., 1994).

In contrast to the SULT2 Northern blot data, which indicated a progressive decline in SULT2 mRNA expression with increasing PB doses, the most demonstrable suppression of immunoreactive SULT2 protein expression occurred at the lowest test dose of PB (SULT2 protein was suppressed to levels that were ~43% of vehicle-treated control levels), an effect that approached statistical significance (p = 0.054). In accordance with the SULT1 mRNA data, which showed minimal changes in hepatic SULT1 mRNA levels in response to PB, amounts of hepatic immunoreactive SULT1 protein were not significantly altered in PB-treated rats (fig. 2).

View larger version (32K): [in this window] [in a new window]   Fig. 2.   Effects of in vivo treatments with PB on immunoreactive SULT1 and SULT2 protein levels in male rat liver. Rats were treated as described in the legend to fig. 1. Cytosol was prepared from a sample of each individual liver, and aliquots (3 µg of protein/lane for SULT1 and 7.5 µg of protein/lane for SULT2) were analyzed in Western blots for amounts of immunoreactive SULT1 (A) or SULT2 (B) protein. Each Western blot contained three control liver cytosol samples (C) and three liver cytosol samples prepared from rats treated with one dose of PB (35, 80, or 100 mg/kg). Immunoreactive proteins were detected by enhanced chemiluminescence, and band intensities were estimated by scanning laser densitometry. The immunoreactive SULT1 and SULT2 bands were of the expected sizes (~35 and ~30 kDa, respectively). The semiquantitative densitometric data are presented as percentages of control (mean ± SD, N = 3).

A recognized limitation of using SULT1 and SULT2 class-specific cDNA probes for Northern blots and polyclonal antisera for Western blots is that these tools lack adequate specificity to detect PB-mediated effects on individual SULT isoforms. Therefore, the impact of PB treatment on the mRNA expression of individual rat hepatic SULT1 and SULT2 isoforms was investigated using isoform-specific oligonucleotide probes. Of the SULT1 isoforms examined, SULT1A1 was the only isoform that showed a significant dose-dependent suppressive response to PB treatment. At PB doses of 80 and 100 mg/kg, SULT1A1 mRNA levels declined significantly, to ~57 and ~42% of control levels, respectively (fig. 3). In contrast, moderate increases in SULT1C1 and SULT1E2 mRNA levels were observed in PB-treated rats, although these changes were not statistically significant (fig. 3). The SULT-Dopa/tyrosine isoform was significantly induced, in a dose-dependent manner. In rats treated with 100 mg/kg PB, hepatic SULT-Dopa/tyrosine mRNA levels were increased significantly (to ~417% of control levels) (fig. 3).

View larger version (22K): [in this window] [in a new window]   Fig. 3.   Effects of in vivo treatments with PB on individual SULT1 isoform mRNA levels in male rat liver. Rats were treated and samples of poly(A)+ RNA were prepared as described in the legend to fig. 1. A, autoradiographs of Northern blots hybridized with SULT1 isoform-specific oligonucleotide probes. Aliquots of the poly(A)+ RNA samples prepared from the three livers comprising each treatment group were pooled, and 2 µg of the pooled samples were analyzed by Northern blot hybridization, using oligonucleotide probes directed against SULT1A1, SULT1C1, SULT1E2, or SULT-Dopa/tyrosine mRNA. A representative autoradiograph of a blot rehybridized with -actin cDNA is also shown. B, graphical representations of SULT1 isoform mRNA band intensities. As described in the legend to fig. 1, aliquots of poly(A)+ RNA (2 µg/lane), prepared from samples of each individual liver, were analyzed by Northern blot hybridization, using isoform-specific SULT1 oligonucleotide probes. Hybridizable RNA band intensities were estimated by scanning laser densitometry. SULT1 isoform mRNA band intensities were normalized to the respective -actin band intensities and are presented as percentages of control (mean ± SD, N = 3). *, significantly different from control (p < 0.05); **, significantly different from control (p < 0.01). * The slope of the best-fit line for the dose-response data is significantly different from 0 (p < 0.05).

Although the SULT2 isoforms are closely related structurally and functionally (Watabe et al., 1994), PB produced markedly different effects on the expression of individual SULT2 isoforms (fig. 4). When mRNA samples from similar treatment groups were pooled, increasing doses of PB clearly produced suppression of SULT20/21, induction of SULT40/41, and a lesser increase in the expression of SULT60 (fig. 4A). Triplicate analysis of separate poly(A)⁺ RNA samples demonstrated dose-dependent decreases in SULT20/21 mRNA expression in response to increasing doses of PB (35, 80, and 100 mg/kg), which were statistically significant by linear regression analysis (to levels that were ~61, ~30, and ~41% of vehicle-treated control levels, respectively) (fig. 4B). In direct contrast, these same doses of PB produced significant (by linear regression) dose-dependent increases in the amounts of hepatic SULT40/41 mRNA (to ~261, ~313, and ~416% of control levels, respectively) and SULT60 mRNA (to ~97, ~133, and ~162% of control levels, respectively) (fig. 4B).

View larger version (23K): [in this window] [in a new window]   Fig. 4.   Effects of in vivo treatments with PB on individual SULT2 isoform mRNA levels in male rat liver. Rats were treated and samples of poly(A)+ RNA were prepared as described in the legend to fig. 1. A, autoradiographs of Northern blots hybridized with SULT2 isoform-specific oligonucleotide probes. Aliquots of the poly(A)+ RNA samples prepared from the three livers comprising each treatment group were pooled, and 2 µg of the pooled samples were analyzed by Northern blot hybridization, using oligonucleotide probes directed against SULT20/21, SULT40/41, or SULT60 mRNA. A representative autoradiograph of a blot rehybridized with -actin cDNA is also shown. B, graphical representations of SULT2 mRNA band intensities. As described in the legend to fig. 1, aliquots of poly(A)+ RNA (2 µg/lane), prepared from samples of each individual liver, were analyzed by Northern blot hybridization, using isoform-specific SULT2 oligonucleotide probes. Hybridizable RNA band intensities were estimated by scanning laser densitometry. SULT2 isoform mRNA band intensities were normalized to the respective -actin band intensities and are presented as percentages of control (mean ± SD, N = 3). * The slope of the best-fit line for the dose-response data is significantly different from 0 (p < 0.05).

	Discussion

Top Abstract Introduction Materials & Methods Results Discussion References

Classes of hormones and xenobiotic agents that induce P450 expression also affect the expression of the SULT multigene family. We previously reported that SULT2 mRNA expression in female rat liver is markedly suppressed in response to in vivo treatment with the CYP1A1-inducing agent 3-methylcholanthrene (Runge-Morris and Wilusz, 1994). Similarly, prototypical glucocorticoid/steroid chemical inducers of CYP3A induce the mRNA expression of both the SULT1 and SULT2 gene subfamilies in vivo and in primary rat hepatocyte cultures (Liu and Klaassen, 1996a; Runge-Morris et al., 1996). The effects of P450 inducers on SULT gene expression appear to be complex. In vivo treatment of male and female rats with steroidal chemicals produces different patterns of effects on the mRNA expression of individual SULT isoforms in rat liver (Liu and Klaassen, 1996a). In this study, we have demonstrated that treatment with the prototypical CYP2B1/2-inducing agent PB also produces differential effects on individual SULT1 and SULT2 isoforms in rat liver.

Isoform-specific effects on drug-metabolizing enzyme gene regulation have the potential to shift the balance of drug metabolism and xenobiotic detoxication profiles in the liver. Despite the considerable substrate-specificity overlap among members of each SULT gene subfamily, isoform-specific differences in SULT substrate preferences have been reported, particularly within the heterogeneous SULT1 gene subfamily. For example, SULT1A1 readily catalyzes the sulfation of minoxidil and a variety of phenols but displays lesser sulfating activity toward the carcinogenic substrate N-hydroxy-2-acetylaminofluorene (Nagata et al., 1993; Yamazoe et al., 1994). In contrast, SULT1C1 shares ~50% amino acid sequence identity with SULT1A1 (Yamazoe et al., 1994) and exhibits robust sulfating activity toward this carcinogenic intermediate (Nagata et al., 1993; Yamazoe et al., 1994). Therefore, xenobiotic agents that preferentially induce or suppress rat hepatic SULT1C1 gene expression may alter the course of hepatocarcinogenesis by interfering with intrahepatic bioactivation or detoxication of this procarcinogen.

EST (SULT1E2) (Demyan et al., 1992) catalyzes the sulfation of estrogenic steroids. However, there appears to be multiplicity within the EST subclass of SULT1 enzymes. To date, five EST cDNA sequences have been published; they appear to represent two distinct isoforms. Comparative sequence analysis of the deduced amino acid sequences for these two closely related isoforms indicates that they share approximately 94% sequence identity (Falany et al., 1995). The SULT1E2 oligonucleotide probe used in the present study was designed to detect the EST isoform described by Demyan et al. (1992) (GenBank accession number M86758) and is a perfect match to the rEST-3 (Falany et al., 1995) (GenBank accession number S76489) and STe1 (Rikke and Roy, 1996) (GenBank accession number U50204) sequences. However, this oligonucleotide almost certainly also hybridizes to mRNA encoding the EST isoform represented by rEST-6 (Falany et al., 1995) (GenBank accession number S76490) and STe2 (Rikke and Roy, 1996) (GenBank accession number U50205), because these EST mRNA sequences and the oligonucleotide probe differ by only one nucleotide. It would be difficult to design any hybridization oligonucleotide capable of discriminating among the closely related EST isoforms, because sequence alignments reveal no ~20-base runs displaying more than three nucleotide differences.

The nucleotide sequence of the SULT-Dopa/tyrosine isoform, which sulfates dopamine and tyrosine residues (Sakakibara et al., 1995; Araki et al., 1997), is almost identical to that of the recently reported SULT1B1 (or ST1B1) (GenBank accession number D89375) (Fujita et al., 1997), which catalyzes the sulfation of 3,3',5'-triiodothyronine hormone (Yamazoe et al., 1994). The coding regions of these two cDNAs differ by only one nucleotide, resulting in an amino acid substitution.

Analysis of the deduced amino acid sequences of SULT2 gene subfamily members suggests that rat and human SULT2 enzymes share ~62% amino acid sequence identity with each other and ~33-39% sequence similarity with the SULT1 gene subfamily members (Yamazoe et al., 1994). In 1989, Ogura et al. screened a rat liver cDNA library with a polyclonal antibody prepared against female-predominant SULT2 (STa) and they cloned ST-20 (SULT20) cDNA (Ogura et al., 1989) and ST-40 (SULT40) cDNA, which shared ~94.4% nucleotide sequence identity with SULT20 cDNA (Ogura et al., 1990a). Subsequently, the SULT21, SULT41, and SULT60 cDNA sequences were identified by rat liver cDNA library screening (Watabe et al., 1994). Comparative analysis of the deduced amino acid sequences for the protein subunits encoded by the rat hepatic SULT2 cDNAs suggests that the SULT20, -21, -40, -41, and -60 deduced protein sequences are structurally very similar, with sequence identities ranging from ~86.3 to ~99.6% (Watabe et al., 1994).

In the liver, the functional significance of microheterogeneity within SULT gene subfamilies is uncertain. For example, SULT20 cDNA differs from SULT21 cDNA by eight nucleotides or six amino acid residues (Watabe et al., 1994), whereas SULT40 and SULT41 differ from each other by only three nucleotide residues or one amino acid (Watabe et al., 1994). To add to the complexity, SULT21 consists of two heterogeneous cDNAs, i.e. ST-21a (SULT21a) and ST-21b (SULT21b), which differ from each other only in the length of their 3'-untranslated regions (Watabe et al., 1994).

As a consequence of opposing effects on individual SULT1 isoforms, the net effects of PB on SULT1 gene subfamily expression, as detected with a SULT1 cDNA probe, were not striking (fig. 1). Similarly, the net results of SULT2 Northern blot and Western blot analyses using a SULT2 cDNA probe and polyclonal antisera, respectively, suggested an overall suppressive effect of PB on SULT2 gene expression. In a previous study of the effects of tamoxifen and PB treatment on drug-metabolizing enzyme gene expression, Northern blot analyses of male rat liver total RNA hybridized with selected SULT cDNA probes suggested that PB increased rSTe1 (SULT1E2) mRNA expression and suppressed rSTa (SULT2) mRNA levels (Hellriegel et al., 1996). The present study establishes, using isoform-specific oligonucleotide probes, that PB treatment produces markedly divergent effects on the mRNA expression of individual SULT isoforms, an effect that may be masked by the use of class-specific cDNA probes in Northern blot analyses.

Despite the strong ~94.4% nucleotide sequence identity shared by the cDNAs corresponding to SULT20 (SULT20/21 isoform) and SULT40 (SULT40/41 isoform) (Ogura et al., 1989, 1990a), PB suppressed SULT20/21 mRNA levels and substantially induced SULT40/41 mRNA expression in rat liver. Like SULT40/41, hepatic SULT60 mRNA levels were increased in response to PB, suggesting that the subtle PB-mediated changes in SULT2 mRNA expression (detected in Northern blots using class-specific SULT2 cDNA probes) conceal more striking and even opposing effects on individual SULT2 isoforms. Noncoordinate pretranslational regulation of the hepatic expression of SULT20/21 and SULT40/41 isoforms in rat liver was previously reported (Ueda et al., 1997). In growth hormone-deficient rats, SULT20/21 (called ST2A1) mRNA levels became undetectable in the livers of both male and female rats, whereas SULT40/41 (called ST2A2) mRNA expression was clearly detectable (Ueda et al., 1997). Delivery of growth hormone by continuous infusion to hypophysectomized rats (female-specific secretory pattern) produced increases in hepatic SULT20/21 mRNA levels but had no significant effects on SULT40/41 expression in male or female rats (Ueda et al., 1997). In contrast, treatment of growth hormone-deficient rats with intermittent injections of growth hormone (male-specific secretory pattern) resulted in a slight increase in hepatic SULT20/21 mRNA expression in mature male rats, no significant effects on SULT20/21 mRNA levels in mature female rats, and marked suppression of hepatic SULT40/41 mRNA expression in both male and female rats (Ueda et al., 1997). These data suggest that gender-dependent growth hormone secretory patterns differentially regulate the hepatic mRNA expression of two highly related SULT2 isoforms (SULT20/21 and SULT40/41).

PB may alter SULT mRNA expression by suppressing or inducing the rates of SULT gene transcription or by modifying SULT mRNA stability. To add to this complexity, it is possible that PB regulates the molecular expression of different SULT isoforms by more than one mechanism, as has been suggested by previous investigations of PB-inducible P450 expression (Burger et al., 1990; Kocarek et al., 1990). Unlike the well-described aryl hydrocarbon receptor-mediated mechanism for the 2,3,7,8-tetrachlorodibenzo-p-dioxin-mediated induction of CYP1A1 gene transcription (Whitlock, 1993), enantioselectivity studies have failed to identify a stereospecific "PB receptor" (James et al., 1981; Nims et al., 1994). However, a 163-base pair fragment of the CYP2B2 5'-flanking region that was found to confer PB responsiveness to a reporter construct transiently transfected into primary cultured rat hepatocytes was recently identified (Trottier et al., 1995). The molecular mechanisms and cellular signaling pathways that mediate the regulation of SULT mRNA expression by PB remain to be elucidated. Future studies addressing this problem will necessarily focus on the ability of PB, structural congeners of PB, and structurally dissimilar, PB-like, CYP2B1/2 inducers to modulate the rates of transcription of individual SULT isoforms in rat liver.