Sulfation and the Sulfotransferases

Top Abstract Sulfation and the... Interindividual Variation and... Regulation of SULT Expression Metabolic Activation of Dietary... Inhibition of Purified... Conclusions References

The sulfation system is a major component of the body's chemical defense armory and is directly involved in the metabolism, detoxification, and elimination of numerous xenobiotics (Coughtrie, 1996; Falany, 1997; Coughtrie et al., 1998). In addition to this important role in xenobiotic metabolism, sulfation also functions in the biosynthesis, mode of action, and homeostasis of many important endogenous chemicals including steroid hormones, iodothyronines, and catecholamines (Visser, 1996; Coughtrie et al., 1998; Eisenhofer et al., 1999). However, for an increasingly large number of xenobiotics, including many environmental pollutants and drugs such as tamoxifen and cyproterone, sulfation results in increased biological activity and for many compounds is the obligatory terminal step in the generation of protein- and/or DNA-adducting species (Glatt, 1997). Variation in sulfation capacity, resulting from environmental insult and/or genetic factors, may be important in determining susceptibility to chemical mutagens. Thus sulfation plays a pivotal role in the balance between toxicity and harmless elimination of a multitude of chemicals, and factors that influence an individual's capacity for sulfation will therefore impact upon these important processes.

Sulfation is catalyzed by sulfotransferase (SULT¹) enzymes encoded by members of the SULT gene superfamily (Weinshilboum et al., 1997). These cytosolic enzymes all utilize 3'-phosphoadenosine 5'-phosphosulfate (PAPS) as the sulfuryl donor (cosubstrate) for the sulfation reaction (Klaassen and Boles, 1997). Humans have at least 11 different SULT enzymes, forming three subfamilies based on their amino acid sequence identity and substrate specificity (Coughtrie et al., 1998) (Table 1, Fig. 1). SULT1 family members (of which seven are currently known and characterized) sulfate primarily phenols (including estrogens and iodothyronines) and catechols (including catecholamines), whereas members of SULT2 family (of which three are known) sulfate primarily steroids, sterols, and other alcohols. A novel SULT, predicted to belong to a separate subfamily, has been identified through the human chromosome 22 sequencing project (Dunham et al., 1999; EMBL accession numbers CAB09788, AF115311). The function of this enzyme is as yet unknown, although phylogenetic analysis suggests it may be (distantly) related to the SULT2 family (Fig. 1B).

View larger version (11K): [in this window] [in a new window]   Fig. 1.   The human sulfotransferase enzyme family. The relationships between different members of the human SULT enzyme family are depicted using amino acid sequence comparisons (A) and phylogenetic analysis (B). A, amino acid sequences for 11 human SULT enzymes were compared using the CLUSTAL W program (Thompson et al., 1994) at the European Bioinformatics Institute web site (http://www2.ebi.ac.uk/clustalw/). Data are shown as percentage of amino acid sequence identity. B, the data generated for panel A were further analyzed and visualized using the TreeView software package (Page, 1996) to generate a rectangular cladogram illustrating the evolutionary relationships within the human SULT enzyme family.

The different SULT isoenzymes share a number of conserved domains, and probably also a similar reaction mechanism. The most striking of these, called region I and region IV (consensus sequences TYPKSGTxW and RKGxxGDWKxxxFT, respectively²) (Weinshilboum et al., 1997) were proposed to be involved in binding of PAPS. The role of these domains in PAPS binding has been confirmed recently following solution of the X-ray crystal structures of mouse estrogen SULT (mEST) (Kakuta et al., 1997, 1998) and the human dopamine SULT, SULT1A3 (Dajani et al., 1999a; Bidwell et al., 1999). A number of other key amino acids in SULT enzymes have also been identified, particularly an absolutely conserved histidine (His₁₀₈ in mEST and SULT1A3), which is proposed to function as the catalytic base in the reaction center (Kakuta et al., 1997). Thus the reaction mechanism as proposed for phenol sulfotransferases would involve an S_N2 nucleophilic attack by the phenoxide (resulting from extraction of the hydrogen atom of the phenol group by His₁₀₈) on the sulfate of PAPS. Other conserved amino acids, including Lys₄₈ and Lys₁₀₆, are proposed to act in stabilizing the reaction intermediate(s) (Kakuta et al., 1998).

A more difficult problem to solve has been understanding what contributes to the specificity of individual sulfotransferases. Although there is considerable overlap in the specificities of a number of the enzymes, they do exhibit distinct selectivities toward certain compounds. For example, SULT1A3 is highly selective for endogenous catecholamines such as dopamine. We recently showed that a single amino acid, Glu₁₄₆, is responsible for this selectivity in SULT1A3 (Dajani et al., 1998, 1999a). Glu₁₄₆ is located at the end of a region of conserved secondary structure in SULT enzymes, -helix 6, and no other SULT identified to date has this amino acid in this position, just as no other SULT has been found to have the high degree of selectivity for dopamine and similar catecholamines. We showed that the conformation of compounds such as dopamine, where a basic nitrogen 2 carbon atoms distant from the aromatic ring forms a salt bridge with Glu₁₄₆, placed the compound in optimum position in the active site of SULT1A3. A substantial amount of work is still required if we are to fully appreciate the molecular mechanisms of sulfotransferase specificity.

	Interindividual Variation and Polymorphisms of Sulfotransferases

Top Abstract Sulfation and the... Interindividual Variation and... Regulation of SULT Expression Metabolic Activation of Dietary... Inhibition of Purified... Conclusions References

Chemical-induced diseases, including cancer, cardiovascular disease and adverse drug reactions, account for substantial mortality and morbidity world-wide (Lazarou et al., 1998). Interindividual differences in the expression and/or activity of drug-metabolizing enzymes is a well established cause of adverse drug reactions and other toxic effects associated with exposure to both xenobiotic and endogenous chemicals, and these differences may arise from a variety (or combination) of genetic and/or environmental events.

There is a relative paucity of information on the polymorphic expression of SULT enzymes, when compared to other xenobiotic-metabolizing systems such as cytochromes P450, glutathione S-transferases, and N-acetyltransferases. The major form of phenol SULT in adult human liver (SULT1A1) is subject to a common genetic polymorphism (Price et al., 1989), and the molecular basis of this has recently been identified (Jones et al., 1993; Coughtrie et al., 1999; Raftogianis et al., 1999). A single mutation in the SULT1A1 gene results in an Arg₂₁₃  His amino acid substitution, which affects the activity and expression of the protein, presumably through reduced protein stability (Raftogianis et al., 1997). Individuals who are homozygous for the SULT1A1*2 genotype (i.e., His₂₁₃/His₂₁₃) have significantly reduced platelet sulfotransferase activity (measured with a SULT1A1-specific substrate) (Raftogianis et al., 1997), and platelet enzyme activity correlates strongly with protein expression (Jones et al., 1993; Raftogianis et al., 1997). We have recently devised a simple polymerase chain reaction-restriction fragment length polymorphism assay for this mutation and have shown that the SULT1A1*2 allele frequency is approximately 32% (Coughtrie et al., 1999), a figure which agrees well with gene sequencing studies (Raftogianis et al., 1999) (Table 2). A number of other mutations in SULT1A1 have been identified but these are present in the population at allele frequencies of less than 1% and do not appear to have functional consequences for the SULT1A1 enzyme (Raftogianis et al., 1999). Allelic variants of another member of the SULT1 family, SULT1A2, are also known from gene sequencing studies (Raftogianis et al., 1999) (Table 2), and the common variant allozyme SULT1A2*2 (allele frequency 29%) displays vastly different kinetic properties to the wild type enzyme, SULT1A2*1. However there is some debate as to the physiological importance of this enzyme, which shares 96% amino acid sequence identity with SULT1A1 (Fig. 1).

Recently, it has been found that the expression of the estrogen sulfotransferase enzyme protein SULT1E1 varies widely between individuals (Song et al., 1998; Rubin et al., 1999). To date, there has been no investigation into the molecular basis of this large interindividual variation in expression at the genetic level. Three cDNA clones encoding human SULT1E1 from different tissues, and the gene located on chromosome 4, have been isolated in different laboratories and the derived amino acid sequences are identical (Aksoy et al., 1994; Falany et al., 1995; Her et al., 1995; Rubin et al., 1999).

	Regulation of SULT Expression

Top Abstract Sulfation and the... Interindividual Variation and... Regulation of SULT Expression Metabolic Activation of Dietary... Inhibition of Purified... Conclusions References

Little is known of the regulation of SULT expression in humans, but it is certainly very different from the situation in rodents, where SULTs exhibit dramatic sexual dimorphism. The complement and tissue distribution of individual isoenzymes also differs considerably between man and laboratory animals, making the extrapolation of animal data to humans difficult, and often of questionable value. There is no evidence for global differences between the sexes in the expression of various SULT isoforms in the liver, although there is obviously gender-dependent expression in reproductive tissues. The genes encoding SULT1A1, -1A2, and -1A3 all show evidence of multiple noncoding 5'-exons, which may be involved in tissue-specific regulation of expression of these proteins (Weinshilboum et al., 1997), although the mechanisms are as yet unknown. However, there is significant tissue-selectivity in the expression of these proteins. For example, in the adult SULT1A3 is essentially not expressed in the liver, whereas it is present at extremely high levels in the upper gastrointestinal tract (Rubin et al., 1996), its principal site of expression, where it plays a key role in the synthesis of circulating catecholamine sulfates (Goldstein et al., 1999; Eisenhofer et al., 1999). In stark contrast, SULT1A3 is expressed at high levels in fetal liver, with the expression being "switched off" at around the time of birth (K. Richard, R. Hume, E. Kaptein, T. J. Visser, and M. W. H. Coughtrie, manuscript submitted).

Steroid SULTs are also subject to tissue-specific regulation. The major steroid-sulfating enzyme, SULT2A1 or dehydroepiandrosterone sulfotransferase, is not expressed at any stage in the endometrium (Rubin et al., 1999), but it is highly expressed in the adrenal gland (more so in the fetus than the adult) (Barker et al., 1994), where it is responsible for production of DHEA sulfate, the major adrenal steroid (Parker and Odell, 1980; Neville and O'Hare, 1982). SULT2A1 is also present at high levels in the liver and intestine.

SULT1E1 is subject to considerable regulatory control by progesterone, and potentially other compounds, in hormone-dependent tissues (e.g., endometrium, breast) (Tseng and Liu, 1981), and this may account for some of the variability in expression observed within the human population. For example, it can be demonstrated in vitro that SULT1E1 expression in endometrium carcinoma cells in culture is induced by progesterone (Falany and Falany, 1996), and this enzyme is obviously under the control of progesterone in vivo as demonstrated by the menstrual cycle-dependent variation in SULT1E1 expression in human endometrium (Buirchell and Hähnel, 1975; Rubin et al., 1999).

Similarly, little is known of the regulation of SULT expression in humans by xenobiotics. In rodents, they have been historically considered refractory to induction by "classical" xenobiotic inducers, although more detailed analysis recently has shown that certain SULT isoforms are responsive to phenobarbital and other compounds, such as peroxisome proliferators and steroidals, with both induction and repression of expression observed depending upon the isoform studied (Liu and Klaassen, 1996; Witzmann et al., 1996; Runge-Morris et al., 1998). Induction of the sulfation of 17-ethinylestradiol and 4-nitrophenol by rifampicin has been demonstrated in human hepatocytes (Kern et al., 1997; Li et al., 1999).

	Metabolic Activation of Dietary Chemicals by Sulfation

Top Abstract Sulfation and the... Interindividual Variation and... Regulation of SULT Expression Metabolic Activation of Dietary... Inhibition of Purified... Conclusions References

The gastrointestinal tract is a major target organ for dietary mutagens and carcinogens, and therefore the presence of substantial amounts of various SULTs in the human gastrointestinal tract is of considerable importance from a toxicological perspective. The well recognized ability of various SULT isoforms to bioactivate numerous xenobiotics, including a host of dietary chemicals, strongly suggests a role for sulfation in the pathogenesis of gastrointestinal tumors. Examples of chemicals that are dependent upon sulfation for generation of DNA- and protein-adducting species include benzylic alcohols of polycyclic aromatic hydrocarbons, estragole and safrole, as well as hydroxyarylamines and arylhydroxamic acids, including those formed from heterocyclic amines found in cooked meat and fish (Meerman and Vandepoll, 1994; Miller, 1994; Chou et al., 1995; Wakazono et al., 1998). There is a considerable body of evidence, from both in vivo and in vitro sources, supporting the role of sulfation and sulfotransferases in these reactions (Miller, 1994; Glatt, 1997; Coughtrie et al., 1998). More recently, studies have focused on determining the role of individual SULT isoenzymes in the bioactivation of dietary and other chemicals (see Glatt et al., 1995a,b, 1996; Glatt, 1997; Wakazono et al., 1998). The major enzymes in humans involved in these bioactivation reactions are SULT1A1, -1A3, -1E1, and -2A1, all of which are known to be expressed in the gastrointestinal tract, including the colon. The interaction of dietary chemicals with sulfation in the gastrointestinal tract may well be important in modulating the activation of dietary pro-mutagens and pro-carcinogens by SULTs in this tissue. Therefore, others and we have been interested in determining whether there are dietary chemicals that act as inhibitors of SULTs, and which may be functioning as natural chemoprotectants.

	Inhibition of Purified Recombinant Human Sulfotransferases by Dietary Chemicals

Top Abstract Sulfation and the... Interindividual Variation and... Regulation of SULT Expression Metabolic Activation of Dietary... Inhibition of Purified... Conclusions References

A number of natural and synthetic dietary chemicals are known to be inhibitors of various human sulfotransferases, including SULT1A1, SULT1A3, and SULT1E1 (Littlewood et al., 1985; Gibb et al., 1987; Bamforth et al., 1993; Jones et al., 1995b; Coughtrie et al., 1998). These include numerous polyphenols such as quercetin (Eaton et al., 1996), components of red wine (Gibb et al., 1987; Jones et al., 1995b), and green tea and coffee (Coughtrie et al., 1998).

We have recently investigated in more detail the selectivity and mechanism of inhibition of human phenol SULTs by various dietary chemicals. These experiments were carried out using Escherichia coli-expressed, purified recombinant SULT1A1, -1A2, and -1A3 prepared as previously described (Dajani et al., 1998, 1999b). First, we chose a series of chemicals that are present in dietary components known to be inhibitors of human phenol sulfotransferases, including coffee, red wine, and tea. The three purified recombinant human SULT enzymes were then used to determine whether any of these compounds were themselves sulfated, and the results are shown in Fig. 2. Of the compounds tested, catechin, resveratrol, and epicatechin were all found to be good substrates for SULT1A3, although activity with the other isoforms was substantially lower. The other compounds demonstrated minimal or no sulfotransferase activity. These compounds were then tested for their ability to inhibit the sulfotransferase activity of SULT1A1, -1A2, and -1A3 (Fig. 3). For these experiments, we used 4-isopropylcatechol as acceptor substrate, because we have found this to be well sulfated by the three sulfotransferase enzymes under investigation (Dajani et al., 1999a; our unpublished work). Figure 3 shows that all the compounds tested inhibited at least one of the SULT enzymes. Epicatechin gallate, epigallocatechin gallate, and gallocatechin gallate strongly inhibited all three enzymes at both concentrations used (10 and 200 µM), whereas other compounds displayed varying degrees of isoform-selectivity. For example, caffeine and chlorogenic acid were the only two compounds to selectively inhibit SULT1A2, whereas quinic acid and trigonelline had no effect on SULT1A1.

View larger version (18K): [in this window] [in a new window]   Fig. 2.   Sulfation of dietary chemicals by human SULT1A enzymes. Various dietary chemicals at concentrations of 10 µM (filled bars) and 200 µM (open bars) were incubated with recombinant human sulfotransferases SULT1A1 (0.5 µg), SULT1A2 (0.35 µg), and SULT1A3 (0.2 µg) and PAP35S (20 µM, 0.02 µCi) in 10 mM phosphate buffer, pH 7.4, for 10 min at 35°C. Sulfotransferase activity was quantified by the barium precipitation method (Jones et al., 1995b) originally described by Foldes and Meek (1973).

View larger version (35K): [in this window] [in a new window]   Fig. 3.   Inhibition of human SULT1A enzymes by dietary chemicals. Sulfotransferase activity was measured using PAP35S and 4-isopropylcatechol as described in the legend to Fig. 2, in the presence and absence of each compound at a concentration of 10 µM. Data show percentage of inhibition of SULT activity toward 4-isopropylcatechol achieved with each compound. The following amounts of purified recombinant enzymes and concentrations of 4-isopropylcatechol were used: 0.5 µg, 18 µM (SULT1A1, filled bars), 0.35 µg, 40 µM (SULT1A2, open bars), and 0.2 µg, 120 µM (SULT1A3, hatched bars).

We further investigated the mechanism of inhibition of SULT1A enzymes by two of the most potent non-substrate inhibitors, epicatechin gallate and epigallocatechin gallate (Fig. 4, A and B, respectively). The results show that although these compounds inhibited all three enzymes, there were significant differences in the mechanism by which they exerted their effects. The two compounds behaved similarly with each enzyme. With SULT1A1, both compounds reduced the K_m and V_max values but had no effect on the efficiency constant V_max/K_m (i.e., uncompetitive inhibition), whereas for SULT1A2 and -1A3 they had no effect on K_m but reduced V_max and V_max/K_m (i.e., mixed type inhibition). Determination of K_i values for both compounds with each of the enzymes demonstrated that their inhibitory potency toward the three enzymes varied considerably, with SULT1A1 being the most susceptible (K_i values of 64 nM and 42 nM with epicatechin gallate and epigallocatechin gallate, respectively) and SULT1A3 the least susceptible (K_i values of 2.8 µM and 1.8 µM, respectively) to inhibition (Table 3).

View larger version (24K): [in this window] [in a new window]   Fig. 4.   Kinetics of inhibition of human SULT1A enzymes by epicatechin gallate and epigallocatechin gallate. Km (µM) and Vmax (nmol/min/mg) were determined, and Vmax/Km ratios calculated for SULT1A1, -1A2, and -1A3 in the presence (filled bars) or absence (open bars) of 2.5 µM epicatechin gallate (A) or epigallocatechin gallate (B) using the PAP35S assay described in the legend to Fig. 2. Reactions contained the following: 0.5 µg SULT1A1 and 0.1 to 30 µM 4-isopropylcatechol; 0.35 µg SULT1A2 and 1 to 100 µM 4-isopropylcatechol; 0.2 µg SULT1A3 and 1 to 100 µM 4-isopropylcatechol.

The results of these studies provide further evidence for the presence of potent inhibitors of phenol sulfotransferases in common dietary components. In addition, they provide novel information on the mechanism of inhibition of SULTs by these compounds and indicate that it may be possible to distinguish the highly similar isoforms SULT1A1 and SULT1A2 on the basis of differential inhibition by compounds such as chlorogenic acid, caffeine, and caffeic acid.

	Conclusions

Top Abstract Sulfation and the... Interindividual Variation and... Regulation of SULT Expression Metabolic Activation of Dietary... Inhibition of Purified... Conclusions References

Sulfotransferases are a complex enzyme family performing a multitude of important functions, most of which are concerned with detoxication of xenobiotics and homeostasis of important endogenous chemicals. However, for certain compounds, the sulfate conjugates are more reactive/toxic than the parent molecule and therefore they are important toxicologically. Considerable variability in SULT expression and/or activity exists within the human population, and this variability is likely to be an important feature in determining a given individual's capacity to sulfate xenobiotics and endogenous chemicals. Many common dietary chemicals are potent sulfotransferase inhibitors, and may act as important modulators of sulfotransferase function. These compounds may also provide us with reagents to study the function of individual sulfotransferase isoenzymes. The powerful molecular tools we now have available are helping us, albeit slowly, to understand the role of sulfation in the normal human being as well as in susceptibility to a variety of chemical-induced diseases.