ToxSci Advance Access originally published online on December 1, 2005
Toxicological Sciences 2006 90(1):5-22; doi:10.1093/toxsci/kfj061
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
REVIEW |
Human Sulfotransferases and Their Role in Chemical Metabolism
,¶
* School of Biomedical Sciences, School of Molecular and Microbial Sciences, and ¶ Institute for Molecular Bioscience, University of Queensland, Queensland 4072, Australia; Department of Medicine/Oncology, Duke University Medical Center, Durham North Carolina 27710; and School of Exercise and Health Sciences, Deakin University, Australia
1 To whom correspondence should be addressed at Faculty of Biological and Chemical Sciences, Computer Science Building (Building 69, Room 304), University of Queensland, Queensland 4072, Australia. Fax: 61-7-3365-1613. E-mail: m.mcmanus{at}uq.edu.au.
Received September 21, 2005; accepted November 28, 2005
 |
ABSTRACT |
|---|
 TOP ABSTRACT INTRODUCTIONSULFOTRANSFERASESLOCALIZATION OF HUMAN...STRUCTURAL STUDIES AND ENZYMATIC...BIOACTIVATIONREGULATION OF SULFOTRANSFERASESCONCLUSIONSUPPLEMENTARY DATAREFERENCES |
|---|
Sulfonation is an important reaction in the metabolism of numerous xenobiotics, drugs, and endogenous compounds. A supergene family of enzymes called sulfotransferases (SULTs) catalyze this reaction. In most cases, the addition of a sulfonate moiety to a compound increases its water solubility and decreases its biological activity. However, many of these enzymes are also capable of bioactivating procarcinogens to reactive electrophiles. In humans three SULT families, SULT1, SULT2, and SULT4, have been identified that contain at least thirteen distinct members. SULTs have a wide tissue distribution and act as a major detoxification enzyme system in adult and the developing human fetus. Nine crystal structures of human cytosolic SULTs have now been determined, and together with site-directed mutagenesis experiments and molecular modeling, we are now beginning to understand the factors that govern distinct but overlapping substrate specificities. These studies have also provided insight into the enzyme kinetics and inhibition characteristics of these enzymes. The regulation of human SULTs remains as one of the least explored areas of research in the field, though there have been some recent advances on the molecular transcription mechanism controlling the individual SULT promoters. Interindividual variation in sulfonation capacity may be important in determining an individual's response to xenobiotics, and recent studies have begun to suggest roles for SULT polymorphism in disease susceptibility. This review aims to provide a summary of our present understanding of the function of human cytosolic sulfotransferases.
Key Words: human sulfotransferases; SULT1A1 regulation; SULT crystal structures; bioactivation.
|
INTRODUCTION |
|---|
|
TOPABSTRACTINTRODUCTIONSULFOTRANSFERASESLOCALIZATION OF HUMAN...STRUCTURAL STUDIES AND ENZYMATIC...BIOACTIVATIONREGULATION OF SULFOTRANSFERASESCONCLUSIONSUPPLEMENTARY DATAREFERENCES |
|---|
Sulfonate conjugation was first reported by Baumann in 1876 (Baumann, 1876
) and has since been shown to be an important pathway in the biotransformation of numerous xeno- and endobiotics such as drugs, chemical carcinogens, hormones, bile acids, neurotransmitters, peptides, and lipids. The universal sulfonate donor for these reactions is 3'-phosphoadenosine 5'-phosphosulfate (PAPS), and the transfer of sulfonate (
) to a hydroxyl or amino- group is catalysed by a super gene family of enzymes called sulfotransferases (SULTs). In the case of most xenobiotics and small endogenous substrates, sulfonation has generally been considered a detoxification pathway leading to more water-soluble products and thereby aiding their excretion via the kidneys or bile. For a drug like acetaminophen or a neurotransmitter such as dopamine, this is certainly the case; however, for xenobiotics such as N-hydroxy arylamines, N-hydroxy heterocyclic amines, and hydroxymethyl polycyclic aromatic hydrocarbons, sulfonation is a metabolic activation process leading to highly reactive electrophiles that are both mutagenic and carcinogenic (Falany, 1997; Weinshilboum et al., 1997). Further, for the hair growth stimulant, minoxidil, and the neuroendocrine peptide cholecystokinin (CCK), the sulfonated forms of these molecules elicit their biological effects (Falany, 1997; Vargas et al., 1994).
Two broad classes of SULTs have been identified: (1) membrane-bound SULTs that are located in the Golgi apparatus of the cell and are responsible for the sulfonation of peptides (e.g., CCK), proteins, lipids, and glycosaminoglycans, affecting both their structural and functional characteristics (Falany, 1997; Negishi et al., 2001) and (2) cytosolic SULTs that are responsible for the metabolism of xenobiotics and small endogenous substrates such as steroids, bile acids, and neurotransmitters. The focus of this review is on the human cytosolic SULTs, with particular emphasis on those isoforms that have been shown to metabolize a broad range of drug, xenobiotic, and endobiotic substrates. Their particular role in the metabolic activation of xenobiotics to mutagens and carcinogens will be addressed.
The sulfonation of xenobiotics and small endogenous substrates such as steroids and neurotransmitters is widely distributed in nature and occurs in organisms ranging from microbes to man (Blanchard et al., 2004; Nagata and Yamazoe, 2000; Rikke and Roy, 1996). The process of sulfonation involves the transfer of a sulfonyl () group, generally to a hydroxyl on an acceptor molecule, which is catalysed by SULTs. On many occasions, this reaction has been incorrectly termed sulfation in the literature, because it creates a sulfated product. Further, the fact that sulfate is the fourth most abundant anion in human plasma (Markovich, 2001) and is used in the synthesis of the universal sulfonate donor, PAPS, has helped many investigators to incorrectly call the process sulfation. The importance of PAPS in the regulation of sulfonation has been expertly reviewed (Klaassen and Boles, 1997; Schwartz, 2005). In general, the overall enzymatic mechanism requires the sulfonate acceptor (ROH) and donor (PAPS) to bind to a sulfotransferase, which results in the release of the sulfonated product and 3'-phosphoadenosine-5'-phosphate (PAP). SULTs are capable of sulfonating a wide range of substrates including phenols (2-naphthol), primary (ethanol) and secondary alcohols (2-butanol), N-hydroxy arylamines (2-acetylaminofluorene; 2-AAF), N-hydroxy heterocyclic amines (2-amino-1-methyl-6-phenylimidazo[4,5-b]pyridine; PhIP), benzylic alcohols of polycyclic aromatic hydrocarbons (1-hydroxymethylpyrene; 1-HMP), phenolic (17ß-estradiol; E2) and alicyclic hydroxysteroids (dehydroepiandrosterone; DHEA), and iodothyronines. SULTs are also capable of sulfonating amino groups of arylamines such as 2-naphthylamine (Jakoby et al., 1980). The broad substrate specificity of SULTs is due to the fact that multiple forms of these enzymes exist and that the binding sites of some isoforms is plastic, allowing the enzyme to adopt varying architectures so that it can interact with small aromatics, L-shaped aromatics, and fused ring compounds (Gamage et al., 2003).
|
SULFOTRANSFERASES |
|---|
|
TOPABSTRACTINTRODUCTIONSULFOTRANSFERASESLOCALIZATION OF HUMAN...STRUCTURAL STUDIES AND ENZYMATIC...BIOACTIVATIONREGULATION OF SULFOTRANSFERASESCONCLUSIONSUPPLEMENTARY DATAREFERENCES |
|---|
Rat SULT2A2 was the first SULT cloned and was originally identified as a senescence marker protein (Chatterjee et al., 1987
). Since then at least 47 mammalian SULT isoforms, one insect isoform, and eight plant enzymes, which represent nine separate SULT families and 14 subfamilies, have been cloned and characterized (Blanchard et al., 2004). The latter paper provides a significant breakthrough in nomenclature of SULTs, which until then was even confusing to those working full-time in the field. It broadly follows the gene family nomenclature systems developed for other drug-metabolizing enzyme families such as cytochromes P450, UDP-glucuronosyltransferase, glutathione transferases, and N-acetyltransferase. The nomenclature system is based on family members sharing at least 45% amino acid sequence identity and subfamily members being at least 60% identical. To date, five distinct gene families of SULTs have been identified in mammals: SULT1, SULT2, SULT3, SULT4, and SULT5 (Blanchard et al., 2004). The SULT3 family has only been found in mouse and rabbit, and has been shown to primarily sulfonate amino groups (Yoshinari et al., 1998a). Sult5a1 has only been isolated from mice, and limited information is available on this family (Nagata and Yamazoe, 2000).
Table 1 summarizes the human forms of SULTs that have been characterized to date. These can be divided into three families and collectively account for thirteen distinct members: SULT1—A1, A2, A3, A4, B1, C2, C4, E1; SULT2—A1 and B1 (SULT2B1_v1 and SULT2B1_v2); SULT4A1 (SULT4A1_v1 and SULT4A1_v2) (Blanchard et al., 2004; Hildebrandt et al., 2004; Mammalian Gene Collection (MGC) Program Team, 2002). At least two SULT genes (human SULT2B1 and SULT4A1) encode two SULT isoforms: SULT2B1—SULT2B1_v1 and SULT2B1_v2 differ in their N-terminal amino acid sequences as a result of either alternate transcription initiation or alternate splicing (Blanchard et al., 2004; Fuda et al., 2002); SULT4A1—isoforms v1 and v2 differ in their C-terminal amino acid sequences (Mammalian Gene Collection (MGC) Program Team, 2002). Figure 1 shows a dendogram of these human SULTs. While SULT1D1 has been isolated from the dog (Tsoi et al., 2001), mouse (Sakakibara et al., 1998a), and rat (Herrmann and Stoffel, unpublished), no equivalent human form of this enzyme has been identified.
|
|
SULT1 Family
SULT1A1.
Members of the SULT1A1 subfamily have been identified in the largest range of species including the rat, mouse, cow, dog, rabbit (refer Blanchard et al., 2004
for references), monkey (Ogura et al. unpublished), pig (Lin et al., 2004), and platypus (Bolton-Grob and McManus, unpublished). A solitary SULT1A1 isoform has been characterized in all the above-mentioned species, but in humans, four SULT1A subfamily members have been identified (SULT1-A1, A2, A3, and A4; Blanchard et al., 2004; Hildebrandt et al., 2004). Their genes are all clustered on the short arm of chromosome 16 and are thought to have arisen following gene duplication or gene duplication plus recombination events (Aksoy et al., 1994; Dooley, 1998a; Rikke and Roy, 1996). The gene sequences of SULT1A1 and SULT1A2 are 93% similar, whereas SULT1A3 shares approximately 60% identity with the other two genes. The differences are most apparent in the 5' promoter and intron sequences, as the coding exons of all three genes are >90% identical. Based on the available data, the human SULT1A1 is most probably the ortholog of the equivalent animal isoforms (Blanchard et al., 2004).
Wilborn et al. (1993) were the first to isolate a SULT1A1 cDNA from a human liver library. The full-length cDNA was shown to encode a protein of 295 amino acids, which had high activity toward the model substrate p-nitrophenol (pNP). The same protein also had activity toward minoxidil as a substrate (Meisheri et al., 1993). The general fidelity of the Wilborn et al. sequence has now been confirmed by a number of groups (Table 1), and allelic variants of SULT1A1 have been shown to exist in the human population. SULT1A1 is by far the dominant SULT1A protein expressed in human liver and has an estimated molecular weight on SDS–PAGE of 32 kDa. This form was initially labelled by Weinshilboum's group as the TS or P (thermostable or phenol, TS-PST) form of sulfotransferase, and SULT1A3 as the TL or M (thermolabile or monoamine, TL-PST) sulfonating form, and collectively they were shown to be responsible for the metabolism of most phenolic compounds (Hempel et al., 2005; Reiter et al., 1983). In relation to their thermostabilities, SULT1A3 shows no activity toward dopamine as a substrate after treatment at 45°C for 15 min, but SULT1A1 retains approximately 90% of its activity toward pNP following similar treatment. It is also possible to differentiate between SULT1A1 and SULT1A3 activity in tissue fractions using the inhibitor 2,6-dichloro-4-nitrophenol (DCNP), as the sulfonation of pNP by the former is highly sensitive to this compound (Veronese et al., 1994). Unlike SULT1A1 or SULT1A3, SULT1A2 exhibits no activity toward dopamine as a substrate and possesses a Km for pNP sulfonation (
70 µM) between that of SULT1A1 and SULT1A3 (Zhu et al., 1996).
SULT1A2.
As indicated above, no equivalent form of SULT1A2 has been identified in species other than humans. Ozawa et al. (1995) were the first to clone a SULT1A2 cDNA from a human liver library, and it was originally given the name ST1A2. These authors showed that COS [cells derived from the kidney of an adult male African green monkey (Cercopithecus aethiops)] cell-expressed SULT1A2 sulfonated pNP, minoxidil, and ß-naphthol, but at a lower rate than SULT1A1 (ST1A3). In another study, Zhu et al. (1996) isolated from a human liver library two forms of this cDNA (HAST4 and HAST4v) that differ by two amino acids (Thr7 to Ile and Thr235 to Asn). In their coding domains HAST4 and HAST4v were 97% and 94% identical to SULT1A1 and SULT1A3, respectively. On expression of these cDNAs in COS cells, the encoded proteins exhibited markedly different Km values for the sulfonation of pNP: Km values for HAST4 and HAST4v being 74 and 8 µM, respectively. However, unlike SULT1A1 or SULT1A3, SULT1A2 exhibits no activity toward dopamine as a substrate, even though it shared >93% amino acid identity with these proteins. At least 13 different allelic variants of human SULT1A2 have been identified that encode four different amino acid changes resulting in six different SULT1A2 allozymes (Raftogianis et al., 1999).
SULT1A3.
To date, a SULT1A3 gene has only been identified in humans, and it appears that through evolutionary pressures we have acquired a SULT1A3 gene whose expressed protein fulfils a specific role in sulfonating catecholamines such as dopamine (Coughtrie, 1998; Dooley, 1998b; Hempel et al., 2005). SULT1A3 was initially isolated from a human brain cDNA library and was called HAST3 (Zhu et al., 1993). The cDNA isolated was shown to encode a 34 kDa protein that was 93% similar to SULT1A1 and under the new nomenclature was termed SULT1A3 (Blanchard et al., 2004, Zhu et al., 1993). Based on its substrate preference for dopamine, thermal stability, and sensitivity to DCNP inhibition, it was shown to be the thermolabile or M-form of sulfotransferase initially identified by Reiter et al. in 1983 (Veronese et al., 1994). The fidelity of the original sequence was subsequently confirmed (Table 1). An identical cDNA was also isolated by Bernier et al. (1994a), who originally reported it as an estrogen sulfotransferase (hEST). A recent study by Hildebrandt et al. (2004) showed that two copies of SULT1A3 exist in the human genome (SULT1A3 and SULT1A4), and they appear to be transcriptionally active. At least four nonsynonymous single nucleotide polymorphisms (cSNPs) were reported by the above authors for these genes, which show different enzyme activity.
SULT1B1.
The first member of this SULT subfamily was cloned from a rat liver cDNA library and was described as the 3,4-dihydroxyphenylalanine (dopa)/tyrosine sulfotransferase, having activity toward tyrosine and dopa (Sakakibara et al., 1995). The same authors also showed that SULT1B1 had activity toward various thyroid hormone substrates, pNP and dopamine. Since then, SULT1B1 enzymes have been isolated from mouse, dog (Blanchard et al., 2004), and brush-tailed possum (Bolton-Grob and McManus, unpublished). The human form of SULT1B1 was isolated and characterized by Fujita et al. (1997) and shown to be the major thyroid hormone sulfotransferase, having slightly higher affinity for the tri-iodothyronine than SULT1A1.
SULT1C.
SULT1C subfamily members have been isolated from a variety of species including Sult1c1, 1c2, and 1c3 from the rat, 1c1 from the mouse, and 1c2 from the rabbit (see Blanchard et al., 2004 for references). Weinshilboum's group was the first to isolate a human 1C2 cDNA from a fetal liver-spleen cDNA library (Her et al., 1997) and demonstrate that the gene was located on chromosome 2 at 2q11.2. The identical construct was also cloned and characterized by Yoshinari et al. (1998b) and Hehonah et al. (1999). Another member of the human 1C subfamily, 1C4, was identified by Sakakibara et al. (1998b). To date, no endogenous substrates have been identified for members of the IC subfamily.
SULT1E.
The SULT1E enzymes have been widely studied due to their important role in steroid homeostasis. The bovine SULT1E1 was the first cDNA cloned as a known sulfotransferase (Nash et al., 1988). Since this initial study, members of the 1E subfamily have been isolated from a number of species including the guinea pig, rat, mouse, and pig (Blanchard et al., 2004). Aksoy et al. (1994) were the first to isolate a SULT1E1 cDNA from a human liver library. The protein encoded by this cDNA was subsequently shown to have high affinity (nM range) for E2 and estrone and a variety of synthetic estrogens, including diethylstilbestrol and tamoxifen (Falany, 1997; Falany et al., 1995). While other SULTs such as SULT1A1 and SULT2A1 exhibit high activity toward E2 and estrone, this only occurs at nonphysiological concentrations (Falany, 1997; Falany et al., 1995). Iodothyronines are also good substrates for SULT1E1 (Kester et al., 1999).
SULT2 Family
The SULT2 family contains the hydroxysteroid sulfotransferases, which have been subdivided into two subfamilies based on their amino acid sequence identities. Generally, all members of the SULT2 family display overlapping substrate specificities toward an array of hydroxysteroids and related compounds (Table 2).
|
SULT2A.
The rat senescence marker protein (SULT2A2) that is predominantly expressed in aging male rats and mentioned above (Chatterjee et al., 1987
) was subsequently shown to be a hydroxysteroid sulfotransferase. This became apparent on the cloning of three additional rat SULT2A isoforms, SULT2-A1, A3, A4 (see Blanchard et al., 2004 for references). SULT2A isoforms have also been identified in a variety of species, with SULT2A1 being isolated from the mouse, rabbit (Blanchard et al., 2004), and monkey (Ogura et al. unpublished). Unlike the rat, which has multiple forms of these 2A enzymes, humans have only a solitary isoform, SULT2A1 (Comer et al., 1993; Forbes et al., 1995; Kong et al., 1992; Otterness et al., 1992). Human SULT2A1 is responsible for the sulfonation of hydroxysteroids including DHEA, androgens, pregnenolone, and bile acids and was initially cloned using liver and fetal adrenal RNA and termed DHEA sulfotransferase for its preferred substrate (Comer et al., 1993; Forbes et al., 1995; Kong et al., 1992; Otterness et al., 1992; Radominska et al., 1990).
SULT2B.
While the SULT2A and SULT2B subfamilies are capable of metabolizing a range of similar substrates, it nonetheless appears that SULT2B subfamily members are predominantly cholesterol sulfotransferases (Javitt et al., 2001). To date, three members of this subfamily have been identified, Sult2b1 from the mouse (Sakakibara et al., 1998a) and two human isoforms (SULT2B1_v1 and v2; Her et al., 1998). Human SULT2B1_v1 is 15 amino acids shorter than SULT2B1_v2 at the amino terminus, which imparts a functional distinction. For example, SULT2B1_v1 preferentially sulfonates pregnenolone, whereas SULT2B1_v2 catalyses the sulfonation of cholesterol (Fuda et al., 2002).
SULT4 Family
Falany's group cloned the first members of the SULT4 family from both human and rat brain cDNA libraries (Falany et al., 2000). These authors termed the cDNAs "brain sulfotransferase-like" (BR-STL), because of their structural similarity to published SULTs and their selective expression in brain tissue. Liyou et al. (2003) cloned an identical cDNA from a human brain cDNA library, and Sakakibara et al. (2002) have identified the equivalent mouse brain isoform. At the amino acid level the human, rat, and mouse isoforms are 97% similar, and based on the Blanchard et al. (2004) nomenclature, they have been classified as SULT4A1. These proteins are orphan enzymes, as no substantial activity toward endogenous or xenobiotic substrates has been demonstrated (Sakakibara et al., 2002). The predominant brain localization of SULT4A1 and the high degree of sequence identity across species is suggestive of an important, yet unidentified physiological function.
|
LOCALIZATION OF HUMAN SULFOTRANSFERASES |
|---|
|
TOPABSTRACTINTRODUCTIONSULFOTRANSFERASESLOCALIZATION OF HUMAN...STRUCTURAL STUDIES AND ENZYMATIC...BIOACTIVATIONREGULATION OF SULFOTRANSFERASESCONCLUSIONSUPPLEMENTARY DATAREFERENCES |
|---|
SULT1A Subfamily
Much of the early work on the specific cellular localization of human SULT1A members is clouded by the fact that it was not until, 1995 that we definitively knew this subfamily consisted of three closely related members that shared >93% identity at the amino acid level (Blanchard et al., 2004
; Hempel et al., 2005). The recent finding of Hildebrandt et al. (2004), which shows that SULT1A3 has undergone a gene duplication and both SULT1A3 and 1A4 appear to be transcriptionally, active requires a more cautious interpretation of SULT1A3 data. However, the available data obtained using an array of methods including metabolic probes, immunohistochemistry, hybridization histochemistry, immunoblotting, and RT-PCR analysis have shown that SULT1A members exhibit probably the widest tissue distribution of any cytosolic SULT subfamily (Blanchard et al., 2004; Dooley et al., 2000; Hempel et al., 2005). SULT1A1 is by far the major adult liver SULT1A subfamily member and has also been identified in brain (Richard et al., 2001; Whittemore et al., 1986; Young et al., 1985), breast (Windmill et al., 1998), intestine (Teubner et al., 1998), endometrium (Falany et al., 1998), adrenal gland, platelets, and placenta (Abenhaim et al., 1981; Hart et al., 1979; Heroux et al., 1989), kidney and lung (Vietri et al., 2003), and jejunum (Sundaram et al., 1989). SULT1A3, as indicated above, is barely detectable in the adult human liver (Eisenhofer et al., 1999; Heroux et al., 1989) but is highly expressed in jejunum and intestine (Eisenhofer et al., 1999; Richard et al., 2001; Sundaram et al., 1989; Teubner et al., 1998) and also present in platelets and placenta (Heroux et al., 1989) and brain (Whittemore et al., 1985; Young et al., 1985). Using hybridization histochemistry, employing a general SULT1A ribo-probe, a specific SULT1A3 ribo-probe, and a SULT1A antibody, a positive signal was observed in epithelial cells lining the lumen of the stomach and the gastric pits, and in the epithelial cells lining the lumen surface of the crypts of Lieberkuhn of the small intestine and colon. In human lung cytosol, SULT1A1 and SULT1A3 proteins are detectable, and histological studies have shown these proteins present in epithelial cells of the respiratory bronchioles (Hempel et al., 2005; Windmill et al., 1998). Since both the intestine and lungs are major portals of entry of drugs and xenobiotics into the body, the above localization pattern suggests that both SULT1A1 and SULT1A3 may play a significant role in the extrahepatic detoxification and metabolic activation of these chemicals. From a developmental perspective, both SULT1A1 and SULT1A3 are abundantly expressed in the fetal liver, but SULT1A3 almost disappears in adult liver and kidney (Cappiello et al., 1991; Hempel et al., 2005; Pacifici et al., 1993; Richard et al., 2001). These results are suggestive of a role for SULT1A members in protecting the fetus from exogenous toxins and in the homeostasis of hormones such as dopamine and iodothyronines. Further, immunoblotting of placenta cytosol showed the presence of both SULT1A1 and SULT1A3, indicating they may have a potential role in the metabolism of xenobiotics entering the fetal circulation from the maternal side (Heroux et al., 1989; Stanley et al., 2001). The fact that placental UDP-glucuronosyltransferases are relatively low and variable in humans (Collier et al., 2002; Pacifici et al., 1998) suggests that SULT1A members may play a significant role in phase II metabolism in this tissue.
The localization and physiological function of SULT1A2 is the least understood of the SULT1A subfamily members. cDNAs of SULT1A2 have been isolated from both human liver and colon libraries (Blanchard et al., 2004; Ozawa et al., 1995; Zhu et al., 1996), and lower mRNA levels than other SULT1A members have been found in liver, kidney, brain, lung, ovary, and some sections of the gastrointestinal tract (Dooley et al., 2000; Glatt et al., 2001). However, it appears that SULT1A2 mRNA expression does not translate into the formation of protein. For example, Nowell et al. (2005), using a specific anti-peptide antibody for SULT1A2, screened more than 200 cytosolic fractions from 10 different human tissues and found no evidence of immunoreactive protein. Dooley et al. (2000) have suggested that the SULT1A2 gene is a quasi-effective pseudogene of SULT1A1 that is occasionally expressed at the RNA level. While SULT1A2 has been shown to be more efficient than SULT1A1 in the metabolic activation of several aromatic amines (Glatt and Meinl, 2004; Meinl et al., 2002) and capable of activating 3-nitrobenzanthrone and its metabolites in model in vitro systems, the current weight of evidence suggests this enzyme is not expressed in vivo in humans. Because of this data, Nowell et al. (2005) have cautioned the interpretation of data from SULT1A2 genotype/disease association studies.
SULT1B Subfamily
The major physiological role for human SULT1B1 appears to be in thyroid hormone metabolism (Fujita et al., 1997; Wang et al., 1998). To date, SULT1B1 (hST1B2) mRNA has been detected in liver, small intestine, colon, and blood leukocytes (Wang et al., 1998). In the same study, it was also demonstrated that SULT1B1 protein was clearly detectable on immunoblots in cytosol from liver, small intestine, and colon.
SULT1C Subfamily
The physiological role of SULT1C members is currently unknown. Her et al. (1997) performed dot blot analysis and obtained positive signal for SULT1C2 (called SULT1C1 in paper) in the adult human stomach, kidney, and thyroid, as well as fetal liver and kidney. Both Yoshinari et al. (1998b) and Hehonah et al. (1999) isolated identical human sequences from fetal liver and adult stomach cDNA libraries, respectively. Sakakibara et al. (1998b) identified another member of this subfamily from human fetal lung that has been termed SULT1C4. At the RNA level, SULT1C4 was shown to be expressed at higher levels in fetal lung and kidney and at lower levels in fetal heart. Positive signal was also found in the adult kidney, ovary, and spinal cord (Sakakibara et al., 1998b).
SULT1E Subfamily
SULT1E1 is present on immunoblots in both human liver and jejunum cytosol (Forbes-Bamforth and Coughtrie, 1994; Falany et al., 1995). Falany et al. (1998) showed that, on immunoblots, SULT1E1 was not detectable in proliferative endometrial cytosol, but was consistently found in the secretory endometrial cytosols. Coughtrie (2002) has also reported significant levels of SULT1E1 in human fetal liver, lung, and kidney, using E2 activity as a diagnostic indicator of this protein.
SULT2A Subfamily
Northern analysis has shown SULT2A1 to be present in the human liver, adrenal, and small intestine (Luu-The et al., 1995; Otterness et al., 1992; Tashiro et al., 2000). In a more extensive investigation using RT-PCR, Javitt et al. (2001) showed that SULT2A1 mRNA was highly expressed in steroidogenic organs (adrenal and ovary), androgen-dependent tissue (prostate), and in the liver, stomach, small intestine, and colon. Barker et al. (1994) have used immunohistochemistry to demonstrate SULT2A1 (DHEA-ST) expression in embryonic human hepatocytes and that this pattern continues into adulthood, when immunostaining is localized around the central vein. The same authors also reported that SULT2A1 expression was detected in fetal zone of the fetal adrenal, and in the adult, staining was localized in the zona reticularis. Further, kidney SULT2A1 immunostaining was present in the proximal and distal tubules, loops of Henle, collecting ducts, and their progenitors.
SULT2B Subfamily
An initial study using Northern analysis showed SULT2B1 to be localized to the human prostate, placenta, small intestine, and trachea (Tashiro et al., 2000). Javitt et al. (2001) extended this study to both isoforms of SULT2B1 using RT-PCR and demonstrated that SUL2B1_v2 was more widely expressed than its counterpart, particularly in a variety of hormone-responsive tissues including the placenta and prostate, which is suggestive of a differential regulatory mechanism of the two transcripts (Geese and Raftogianis, 2001). For example, SULT2B1_v2 mRNA was present in adrenal gland, placenta, ovary, prostate, lung, kidney, colon, stomach, small intestine, spleen, thymus, thyroid, and liver, whereas SULT2B1_v1 mRNA was present in all these tissues except the last six. Interestingly, SULT2A1 mRNA is present in brain and bone marrow but absent from the skin, where the mRNA of both isoforms of SULT2B are expressed.
SULT4A Subfamily
The current data shows an exclusive brain localization of SULT4A1 in rats, mice, and humans (Falany et al., 2000; Liyou et al., 2003; Sakakibara et al., 2002). However, recent microarray data is suggestive of a wider tissue distribution than the brain for the localization of this SULT (Mm.248796 at <http://www.ncbi.nlm.nih.gov/Unigene/>).
|
STRUCTURAL STUDIES AND ENZYMATIC MECHANISM IN CYTOSOLIC SULTS |
|---|
|
TOPABSTRACTINTRODUCTIONSULFOTRANSFERASESLOCALIZATION OF HUMAN...STRUCTURAL STUDIES AND ENZYMATIC...BIOACTIVATIONREGULATION OF SULFOTRANSFERASESCONCLUSIONSUPPLEMENTARY DATAREFERENCES |
|---|
SULTs can exhibit quite broad, overlapping substrate specificities; however, individual enzymes often demonstrate strict regio-specificity toward a particular substrate (Falany, 1997). Understanding the structural basis of such specificity has been crucial for elucidating the catalytic mechanism and function of these enzymes. It will also aid in predicting the metabolic fate of drugs and chemical carcinogens that are sulfonated and could provide a more rational approach to drug design and chemical risk assessment.
The major catecholamine SULT, SULT1A3, was the first human cytosolic SULT to be structurally characterized (Bidwell et al., 1999; Dajani et al., 1999a). Prior to this, the mouse cytosolic estrogen SULT (SULT1E1) (Kakuta et al., 1997) and the SULT domain of the human Golgi membrane-bound heparan sulfate N-deacetylase/N-sulfotransferase-1 (HSNST; Kakuta et al., 1999) were published. Since then, the number of crystal structures of cytosolic and membrane-bound SULTs have grown rapidly. Despite this progress, only a few crystal structures have been determined with both cofactor and substrate bound (Table 3). This relative lack of information has hindered elucidation of the structural principles underlying the recognition and utilization of a given substrate.
|
Overall SULT Structure
The SULT crystal structures have shown that the enzymes are generally globular proteins with a single
/ß domain that forms characteristic five-stranded parallel ß-sheet surrounded on either side with -helices (Fig. 2). The ß-sheets contribute the PAPS-binding site and the core catalytic residues. These catalytic residues have been shown to be conserved across the cytosolic and membrane-bound SULTs. Interestingly, SULTs share similarities in their structural fold with nucleotide kinases such as uridylate kinase, adenylate kinase, and guanylate kinase (Kakuta et al., 1997).
|
PAP Binding Site
Early sequence analysis suggested that the sequence motif GxxGxxK, present in nearly all cytosolic SULTs, and shown to be important for PAPS binding, was equivalent to the P-loop found in ATP- and GTP-binding proteins (Chiba et al., 1995; Driscoll et al., 1995). However, structural studies on mouse SULT1E1 revealed that the "P-loop" equivalent in SULTs actually corresponds to another loop region, comprising residues 45-TYPKSGT-51 of SULT1E1 (Kakuta et al., 1997
). This loop, termed the phosphosulfate binding (PSB) loop, of SULTs provides the major binding site for the 5'-phosphate group of PAP (Kakuta et al., 1997) and is thought to be important for orienting the cofactor for in-line sulfuryl transfer to the acceptor substrate (Dajani et al., 1999a; Kakuta et al., 1997; Pedersen et al., 2002). Please note that the term PAP is used because mouse SULT1E1 was crystallized in the presence of this molecule and not with PAPS.
The 3'-phosphate of PAP interacts with two conserved regions of sequence, residues 257–259 located at the beginning of the GxxGxxK region, and two additional residues, Arg130 and Ser138 in mouse SULT1E1 (Kakuta et al., 1997). The positioning of the adenine ring of PAP is determined by the residues Trp53, Thr227, and Phe229. These general features identified in SULT1E1 for binding PAP are conserved in other SULT structures, suggesting that the cytosolic and membrane-bound SULTs evolved from a common ancestor (Yoshinari et al., 2001).
Substrate Binding Site
Generally, cytosolic SULTs have a covered hydrophobic substrate binding site, whereas the presumed substrate binding pocket of the membrane-bound HSNST is a large open cleft to allow sulfonation of carbohydrates, glucosaminylglycans, and proteins. As mentioned before, SULTs display broad substrate specificity, though a given enzyme can often be characterized by having a preference for a specific substrate. The underlying principles that regulate this specificity most probably reside in the substrate binding sites of these enzymes. Thus, in contrast to the PAPS binding site, which is characterized by conserved residues across all the SULTs, the substrate binding pocket of SULTs shows a great deal of variability. Despite the elucidation of several SULT crystal structures, the structural principles that underpin substrate specificity are still not fully understood. As indicated above, this is due to the fact that only a few crystal structures have been solved with bound substrate and cofactor present.
In the SULT1A family, the crystal structures of SULT1A1*2, SULT1A1*3, SULT1A2, and SULT1A3 (Table 3) have been determined. The Bidwell et al. (1999) structure of SULT1A3 was solved with a sulfate ion in the cofactor site, whereas the Dajani et al. (1999a) SULT1A3 structure was complexed with PAP. Both structures show large stretches of disordered regions, and this was thought to be a consequence of the lack of a bound substrate. In contrast, the crystal structure of SULT1A1 was crystallized in the presence of both PAP and a model xenobiotic substrate pNP that revealed an L-shaped and very hydrophobic substrate-binding pocket (Gamage et al., 2003). Indeed, Gamage et al. (2003) has shown that the binding site of SULT1A1 is plastic, allowing this enzyme to adopt varying architectures so that it can interact with small aromatics (pNP), L-shaped aromatics (diiodothyronine), and fused ring compounds (E2; Gamage et al., 2005).
Human SULT1A1 and 1A3 share 90% sequence identity, though they exhibit distinct substrate preferences. SULT1A1 prefers uncharged simple phenolic compounds such as pNP, p-cresol, or p-ethylphenol, whereas 1A3 prefers positively charged substrates such as dopamine or tyramine (Brix et al., 1999b). The crystal structure of SULT1A1 revealed the hydrophobic substrate binding pocket mentioned above, which clearly favors binding of uncharged substrates. By contrast, the SULT1A3 substrate binding site includes acidic residues such as Glu146 and Glu89, which favors binding of positively charged substrates. Indeed, site-directed mutagenesis and molecular modeling studies identified Glu146 as a critical residue for the recognition of dopamine by SULT1A3 (Brix et al., 1999a ; Dajani et al., 1998).
Mouse SULT1E1 was the first cytosolic SULT to be structurally characterized (Kakuta et al., 1997); its structure showed the enzyme in a complex with both the substrate E2 and PAP. In the active site, His108 is directly coordinated to the 3-phenolic group of E2 and acts as the catalytic base in the sulfuryl transfer mechanism. This histidine residue is conserved in all cytosolic SULTs, and its mutation has been shown to abolish activity of mouse SULT1E1 (Kakuta et al., 1998). Thus, this structure provided the basis for understanding ß-estradiol binding in the active site and for catalysis of sulfonation by the proposed SN2 in-line displacement mechanism. The crystal structures of SULT1E1:PAP:vanadate (Kakuta et al., 1998) and human SULT1E1:PAPS (Pedersen et al., 2002) provided further evidence as to the structure of the transition state during sulfuryl transfer and gave supporting evidence for the proposed mechanism.
In the SULT2A subfamily, the crystal structures of the human dehydroepiandrosterone sulfotransferase enzymes (SULT2A1, SULT2A3), which sulfonate steroids such as DHEA, androsterone, E2, and pregnenolone, have been solved in complex with PAP (Pedersen et al., 2000), DHEA (Rehse et al., 2002), and androsterone (Chang et al., 2003). In the DHEA-bound structure, two alternative substrate-binding orientations were identified for DHEA, and the authors suggested that the second orientation may reflect a binding mode associated with substrate inhibition. The work of Chang et al. (2003) demonstrated that this enzyme recognizes androsterone as a cognate substrate, with similar kinetics but higher specificity and stronger substrate inhibition than DHEA.
In the SULT2B subfamily, SULT2B1_v1 (SULT2B1a) and 2B1_v2 (2B1b), as outlined above, are splice variants (Her et al., 1998) and have different substrate specificities (Fuda et al., 2002). The crystal structures of SULT2B1_v1 and 2B1_v2 bound with PAP and that of SULT2B1_v2 with its substrate pregnenolone have been determined (Lee et al., 2003). These structures reveal a different catalytic binding orientation for the acceptor substrate, pregnenolone, than that observed for the related steroid DHEA in SULT2A1. It was shown that the amino-terminal helix comprising residues19–26 determines the substrate specificity between the two isoforms. The residues 19-Asp-Ile-Ser-Glu-Ile-23 are responsible for the ability of SULT2b1_v2 to sulfonate cholesterol (Fuda et al., 2002; Lee et al., 2003). Thus, the substrate specificity difference between the two SULT2B1 isoforms appears to lie at the unique amino terminus.
The crystal structures of SULT1B1:PAP, SULT1C1:PAP, and SULT4A1 have also been determined and submitted to the Protein Data bank (PDB, http://www.rcsb.org/pdb), but are not yet described in the literature (Table 3).
Sulfuryl Transfer Mechanism
Duffel and Jakoby (1981) reported that pNP sulfonation by rat aryl sulfotransferase IV has a random Bi Bi mechanism in which PAPS and pNP bind to the enzyme independently. Kinetic studies on the catalytic mechanism of recombinant mouse SULT1E1 suggested that sulfonation follows a random Bi Bi mechanism with dead-end complexes (Zhang et al., 1998). On the other hand, studies using purified human brain aryl sulfotransferase (Whittemore et al., 1986) and flavonol sulfotransferase (Varin and Ibrahim, 1992) suggested that sulfonation occurs via an ordered Bi Bi mechanism. All these studies agree that the sulfonate transfer reaction occurs without formation of intermediates. When the crystal structure of mouse SULT1E1 was solved in the presence of PAP and E2, it became clear that the core structure resembles that of uridylate kinase, with striking similarities between the PAP and ADP binding sites (Kakuta et al., 1997). These structural features of SULT1E1 suggest that the sulfotransferase reaction takes place via an SN2 in-line displacement, a mechanism similar to phosphoryl transfer (Kakuta et al., 1997, 1998). The active site and transition state mimicked by SULT1E1:PAP:vanadate (Kakuta et al., 1998) and human SULT1E1:PAPS complexes (Pedersen et al., 2002) provide further supporting evidence for this mechanism.
Substrate Inhibition
Substrate inhibition, observed at high concentrations of their preferred substrates, is a characteristic feature of SULTs (Raftogianis et al., 1999; Reiter et al., 1983). However, several of the published studies have assumed a Michaelis–Menten model to analyse the kinetics of these enzymes using limited substrate concentration ranges below the overtly inhibitory range (Brix et al., 1999b; Lewis et al., 1996). The recent studies on the crystal structure of SULT1A1 from our laboratory gave the first clues to the molecular basis of substrate inhibition that takes place with small planar substrates such as pNP (Gamage et al., 2003). In the SULT1A1:PAP:pNP structure (Gamage et al., 2003), we observed two pNP molecules bound in the L-shaped active site. When the kinetic implications of this observation were investigated using a wide array of pNP concentrations, we found that there was slight positive cooperativity at low substrate concentrations and substrate inhibition at higher pNP concentrations (above 2 µM; Fig. 3A). From these data, a general kinetic model was constructed (Fig. 3B). The model fits well to the experimental data, and we proposed that impeded catalysis results when both binding sites are occupied, and this gives rise to the observed substrate inhibition with pNP in SULT1A1. Furthermore, we have also shown by molecular modeling and site-directed mutagenesis that the SULT1A3 active site could accommodate two molecules of dopamine (Barnett et al., 2004). From these studies, we have been able to conclude that the substrate inhibition at high concentrations of the substrate is due to impeded catalysis when both binding sites are occupied.
|
The mechanism of substrate inhibition that takes place with multi-ring substrates such as E2 is not well understood. SULT1A1 and SULT1E1 both show strong substrate inhibition at high substrate concentrations of E2 (Adjei and Weinshilboum, 2002
; Falany and Falany, 1997; Fig. 4A). In our recent crystal structure of SULT1A1, there is one molecule each of PAP and E2 in the active site, but the latter is bound in a nonproductive mode (Gamage et al., 2005). This has led us to propose a model (Fig. 4B) that gives an excellent quantitative explanation of the observed substrate inhibition by E2. In this model, a dead-end complex is formed during catalysis, as deduced previously from kinetic studies (Duffel and Jakoby, 1981; Yang et al., 1998). The SULT:PAP:E2 complex that we crystallized is a direct demonstration of this type of dead-end complex, and it provides a ready explanation for the substrate inhibition. A similar type of model may also explain the substrate inhibition that is observed for other SULTs.
|
Dimerization
Most of the cytosolic SULTs generally exist as dimers in solution, and it appears that they are capable of forming not only homodimers but also heterodimers (Petrotchenko et al., 2001
). A conserved dimerization motif was identified by Petrotchenko et al. (2001) in human SULT1E1 consisting of 10 residues near the C-terminus and represented by the consensus sequence KXXXTVXXXE (the so-called KTVE motif). The KTVE motif is conserved in nearly all SULTs (Petrotchenko et al., 2001), though the physiological significance of dimerization for the function of SULTs has not yet been identified. However, biophysical studies carried on dimerization and activation of epidermal growth factor receptor (EGFR) by ligand binding may provide future guidance in understanding the role of dimerization in SULT activity (Schlessinger, 2002).
Molecular Modeling of N-Hydroxy Metabolites of 2AAF and PhIP into SULT1A Isoforms
At the time of writing this review, no structural studies have been published that demonstrate the binding of N-hydroxy aromatic and heterocyclic amines or hydroxy methyl polycyclic aromatic hydrocarbons to human SULTs. However, using the published structures of SULT1A1 (Gamage et al., 2003), SULT1A3 (Bidwell et al., 1999), and the computer model of SULT1A2, we have investigated their ability to accommodate these carcinogens.
Specifically, we have modeled the binding of the N-hydroxylated metabolites of the model carcinogen 2-AAF and the major food-derived mutagen PhIP into the active sites of the three enzymes to investigate their binding interactions. The structures of SULT1A2*1 (unpublished data) and 1A3 (Barnett et al., 2004) were modeled based on the crystal structure of SULT1A1 (Gamage et al., 2003). PAPS was modeled into each of the three structures, based on the crystal structure of the SULT1E1:PAPS complex (Pedersen et al., 2002). The two ligands were docked into each of the three enzyme structures using the GOLD software (Jones et al., 1997, Fig. 5). The allozyme of SULT1A2 (SULT1A2*1 (HAST4v)) and 1A3 we chose to model with is the wild type. The crystal structure of SULT1A1 is that of SULT1A1*2, which has similar binding characteristics to wildtype SULT1A1*1 (with respect to PAPS and pNP) and has slightly lower thermal stability (Rafogianis et al., 1999).
|
Our results showed that, in all three enzymes, the hydroxyl group of N-OH-2-AAF is within hydrogen bonding distance of the donor sulfonate group of PAPS and of the catalytic residue H108 (Fig. 5), consistent with the possibility of catalysis occurring. In the SULT1A1 and 1A2 models, residues F84 and F76 form stacking interactions with the ligand (Figs. 5A and 5B). However, in the 1A1 model the side chain of F81, the gate residue at the active site, forms an unfavorable interaction with the ligand (2.2 Å). In this case, it seems that a conformational change is necessary in F81 to accommodate the ligand. This is not seen in the SULT1A2 model. This observation could perhaps explain the higher activation of N-OH-2-AAF by SULT1A2 compared with SULT1A1 (Glatt, 2000
; Meinl et al., 2002). In our SULT1A3 model (Fig. 5C), the ligand makes unfavorable interactions with both F81 (2.8Å; gate residue) and V84 (2.9 Å) at the enzyme active site. In addition, the active site is relatively acidic as a consequence of residues including E146 and E89; such an environment would be less favorable than the uncharged active sites of SULT1A1 and SULT1A2 for the binding of hydrophobic substrates. Once again, this finding is consistent with the reported lower activation of N-OH-2-AAF by SULT1A3 (Glatt, 2000; Meinl et al., 2002).
The modeling of N-OH-PhIP into the active sites of the enzymes revealed catalytically competent orientations for binding to all three SULT1A structures. In the SULT1A1 and 1A2 models, F76 and F84 form favorable stacking interactions with the benzyl ring of PhIP, as shown in Figures 6A and 6B. Positioning of the residues in the SULT1A1 and 1A2 active sites are similar except for Y149 in SULT1A2. This residue is a histidine in SULT1A1. In the SULT1A2 model, the phenolic hydroxyl of Y149 forms an interaction with the nitrogen of the N-OH group of PhIP. It is not clear whether such an interaction is favorable or unfavorable for binding and catalysis, but given that the interaction is absent in the SULT1A1:N-OH-PhIP model and the two models are otherwise similar, this specific interaction could perhaps explain the lowered metabolic activation of PhIP by SULT1A2 compared with SULT1A1 (Ozawa et al., 1994). By comparison with the very hydrophobic active sites of SULT1A1 and SULT1A2, the SULT1A3 active site is highly charged (Brix et al., 1999a; Dajani et al., 1999a). In our SULT1A3:PAPS:N-OH-PhIP model, we find that the acidic residues (E146 and D86) in the SULT1A3 active site form unfavorable interactions with N-OH-PhIP (Fig. 6C). This could, at least in part, explain the lowered activation of this compound observed with SULT1A3 (Glatt, 2000).
|
Human SULT1E1 and Hydroxylated Polychlorinated Biphenyls (OH-PCBs)
It is reported that certain environmentally relevant PCBs (polychlorinated biphenyls) such as 4,4'-OH-3,5,3',5'-tetraCB inhibit human SULT1E1 at subnanomolar concentrations (Kester et al., 2000). PCBs are man-made pollutants that persist in the environment and exert a variety of toxic effects on experimental animals by inducing estrogenic activities through increased availability of estradiol (Kester et al., 2000). These compounds and their metabolites have been shown to cause endocrine-disrupting effects such as disturbance of sexual development and reproductive function in animals and humans (Cheek et al., 1998). To address the structural basis of inhibition, the crystal structure of human SULT1E1 was determined with the above environmental pollutant and the cofactor product PAP (Shevtsov et al., 2003
). The structure reveals that this compound binds in the active site of SULT1E1 in a very similar way to estradiol (E2) but with a 30° twist between the phenyl rings. This suggests that certain OH-PCBs can act as competitive inhibitors by mimicking E2 binding to SULT1E1 (Shevtsov et al., 2003). |
BIOACTIVATION |
|---|
|
TOPABSTRACTINTRODUCTIONSULFOTRANSFERASESLOCALIZATION OF HUMAN...STRUCTURAL STUDIES AND ENZYMATIC...BIOACTIVATIONREGULATION OF SULFOTRANSFERASESCONCLUSIONSUPPLEMENTARY DATAREFERENCES |
|---|
The early work of the Millers (Millers, 1970, 1978) on the model carcinogen 2-AAF plus other aromatic amines led to the hypothesis that most chemicals require metabolism before being mutagenic or carcinogenic. The first step in the activation of AAF is CYP1A2-mediated N-hydroxylation to N-hydroxy-2-acetylaminofluorene, which is then a substrate for sulfonation, N,O-acyltransfer or deacetylation (McManus et al., 1984
; Thorgeirsson et al., 1983). The influential studies by King and Phillips (1968) and the Millers (DeBaun et al., 1968, 1970) were the first to demonstrate the importance of sulfonation as one of the pathways involved in the activation of 2-AAF. These studies followed the seminal finding of Cramer et al. (1960), which showed that AAF undergoes N-hydroxylation to N-hydroxy-AAF. This metabolite was found to be more carcinogenic than the parent amide, was often active locally, and was active in the guinea pig that had been shown to be resistant to AAF-induced carcinogenicity (Miller, 1970; Miller, 1978). Sulfonation has now been shown to be important in the activation of a range of compounds such as aminoazo dyes, benzidines, heterocyclic amines, hydroxymethyl polycyclic aromatic hydrocarbons, terpenes, ß-aminoethyl alcohols, and 2-nitropropane (Michejda et al., 1994). Recently, this area has been the subject of a range of reviews (Banoglu, 2000; Glatt, 1997, 2000, 2005; Glatt et al., 1998, 2001; Kauffman, 2004; Surh, 1998), and much of the focus has been on determining the specificity of individual sulfotransferase isoforms for the above substrates using model in vitro systems. These studies have shown that certain sulfotransferases are very selective in their activation of promutagens (Glatt, 1997, 2000, 2005), and since they also display tissue-specific expression, it is highly likely that these factors play a role in the organ-selective toxicity of their substrates (Glatt, 2001; Thorgeirsson et al., 1983).
Polymorphisms have been shown in a number of sulfotransferases, and this area has been expertly reviewed by a number of investigators (Glatt, 2000; Glatt and Meinl, 2004; Meinl et al., 2002; Weinshilboum and Adjei, 2005). It was shown that the human SULT1A*Arg (*1) alloenzyme expressed in Salmonella typhimurium cells was 12–350 times more active in metabolizing 2-nitropropane, 2,4-dinitrobenzylalcohol, 1-HMP, (–)-1-(-hydroxyethyl) pyrene, and 2-acetylamino-4-hydroxylaminotoluene to mutagens compared to cells expressing SULT1A1*His (Glatt, 2000). Enantioselectivity has also been demonstrated, with human SULT1E1 and SULT2A1 exhibiting 160-fold preference for the (–)-enantiomer and a 15-fold preference for the (+) enantiomer of 1-(-hydroxyethyl)pyrene, respectively. Our understanding of such differences in the ability of different sulfotransferases to metabolize chemicals has recently been enhanced with the resolution of a number of the crystal structures of these enzymes (Table 3).
|
REGULATION OF SULFOTRANSFERASES |
|---|
|
TOPABSTRACTINTRODUCTIONSULFOTRANSFERASESLOCALIZATION OF HUMAN...STRUCTURAL STUDIES AND ENZYMATIC...BIOACTIVATIONREGULATION OF SULFOTRANSFERASESCONCLUSIONSUPP |
|---|