Sulfation Through the Looking Glass—Recent Advances in Sulfotransferase Research for the Curious

Sulfation through the looking glass—recent advances in sulfotransferase research for the curious

MWH Coughtrie ABSTRACT Members of the cytosolic sulfotransferase (SULT) superfamily catalyse the Department of Molecular & Cellular Pathology, sulfation of a multitude of xenobiotics, hormones and neurotransmitters. University of Dundee, Ninewells Hospital & Humans have at least 10 functional SULT genes, and a number of recent Medical School, Dundee, Scotland, UK advances reviewed here have furthered our understanding of SULT function. Correspondence: Analysis of expression patterns has shown that sulfotransferases are highly MWH Coughtrie, Department of expressed in the fetus, and SULTs may in fact be a major detoxification Molecular & Cellular Pathology, University enzyme system in the developing human. The X-ray crystal structures of of Dundee, Ninewells Hospital and three SULTs have been solved and combined with mutagenesis experiments Medical School, Dundee DD1 9SY, Scotland, UK. and molecular modelling, they have provided the first clues as to the factors Tel: +44 (0)1382 632510 that govern the unique substrate specificities of some of these enzymes. In Fax: +44 (0)1382 640320 the future these and other studies will facilitate prediction of the fate of E-mail: [email protected] chemicals metabolised by sulfation. Variation in sulfation capacity may be important in determining an individual’s response to xenobiotics, and there has been an explosion in information on sulfotransferase polymorphisms and their functional consequences, including the influence of SULT1A1 genotype on susceptibility to colorectal and breast cancer. Finally, the first gene knockout experiments with SULTs have recently been described, with the generation of estrogen sulfotransferase deficient mice in which reproductive capacity is compromised. Our improved understanding of these enzymes will have significant benefits in such diverse areas as drug design and development, cancer susceptibility, reproduction and development. The Pharmacogenomics Journal (2002) 2, 297–308. doi:10.1038/sj.tpj.6500117

Keywords: sulfotransferase; sulfation; pharmacogenetics; structural biology; development

Sulfation of a vast array of natural and synthetic chemicals as well as many biomolecules occurs widely in nature, from bacteria to humans. This article concerns the formation of sulfate conjugates of xenobiotic and endogenous (endobiotic) small molecules by members of the cytosolic sulfotransferase (SULT) enzyme family. The process was first discovered in the late 1870s by Eugen Baumann, who isolated and characterised the sulfate conjugate of phenol from urine of a patient treated with carbolic acid as an antiseptic. It would be another 80 years, however, before the mechanistic basis of sulfate conjugation could be identified, with the discovery of so-called ‘active sulfate’ by Lipmann’s group.1 This compound, that we now call PAPS (30-phosphoadenosine 50-phosphosulfate), is the sulfuryl donor for the vast majority of sulfation reactions2 (Box 1). Recent research in this field that will be highlighted here, has uncovered some important aspects of the function of sulfation. The relative importance of sulfation in humans has perhaps been underestimated in the past, and one of the main reasons for this is that there are major inter-species differences in the function of sulfation—in particular relating to its role in modulating the function of Received: 28 February 2002 Revised: 28 March 2002 hormones and neurotransmitters. Importantly, catecholamines, estrogens, Accepted: 4 April 2002 iodothyronines and dehydroepiandrosterone (DHEA) exist in the human Recent advances in sulfotransferase research MWH Coughtrie 298

Box 1 - Sulfotransferase Reaction and Co-substrate Synthesis

Paracetamol Paracetamol Sulfate

NH CH NH CH3 3 Sulfotransferase O O HO HO 3S 2 APS PAPS PAP

ATP 1 ATP ADP PPi

SO4-- The sulfotransferase reaction relies on the co-substrate PAPS as sulfuryl donor. PAPS is synthesised from inorganic sulfate and 2 molecules of ATP - it is therefore an "expensive" commodity for cells to produce, and is present at relatively low concentrations, typically no more than 20 or 30µM. Low PAPS levels and rapid depletion is believed to limit the capacity of sulfation compared to e.g. glucuronidation, where the co-substrate (UDP-glucuronic acid) is in more plentiful supply. In animals, synthesis of PAPS is the responsibility of a single bifunctional enzyme called PAPS synthetase (PAPSS) that contains both the ATP sulfurylase (1) and APS kinase (2) catalytic activities. In humans and mice, 2 isoforms of PAPSS exist that are the products of the PAPSS1 and PAPSS2 genes. Mutations in PAPSS2 are responsible for rare inherited connective tissue disorders in mouse (brachymorphism) and humans (spondyloepimetaphyseal dysplasia). It is postulated that interindividual variation in the ability to synthesise PAPS, arising from genetic and/or environmental effects on PAPSS, may influence sulfation capacity. circulation predominantly or significantly as the sulfated SULTs related to the mammalian cytosolic enzymes, that are form, whereas this is not the case for rodents or many other involved in flavonol and brassinosteroid metabolism.16 ‘experimental’ animals.3 This is reflected in the different Other model organisms whose genomes have been (or are SULT isoenzyme profiles that exist in various species. For being) sequenced possess various SULTs—the zebrafish example, to date human is the only species in which a (Danio rerio) and amphibian (Xenopus laevis, Silurana tropi- catecholamine-specific SULT isoform (SULT1A3) has been calis) sequence databases contain numerous cytosolic SULT identified, consistent with the high circulating levels of sequences with orthologs in mammalian species, although catecholamine sulfates found in humans and other pri- the Drosophila melanogaster, Caenorhabditis elegans and mates.3,4 Rodents have multiple isoforms of the hydroxyster- Saccharomyces cerevisiae databanks are almost devoid of such oid SULT subfamily 2A,5–7 whereas so far only a single sequences. A new nomenclature system for the cytosolic SULT2A enzyme has been identified and characterised in SULTs has recently been developed* which will (hopefully) humans.8 Also, it is now clear that SULTs are abundantly help to clarify the relationships between the various SULTs expressed in the human fetus9–13—other ‘drug metabolising for novices and aficionados alike. The nomenclature used enzymes’ such as the cytochromes P450 (CYPs), UDP- here follows this new system. glucuronosyltransferases (UGTs) etc are, in general not The SULT1 and SULT2 families are the largest, and expressed at significant levels until after birth14,15—thus probably the most important for xenobiotic and endobiotic SULTs may represent a front line of chemical defence in the metabolism. At present the human SULT enzyme family, developing human. which is the best characterised, comprises of 11 isoforms This article aims to provide the reader with an up-to-date representing the SULT1, SULT2 and SULT4 families. These overview of sulfation in humans, and to highlight some enzymes are the products of 10 genes (SULTs 2B1a and 2B1b important recent research that has impacted significantly on are generated by alternate splicing of the first exon of our knowledge and understanding of this important meta- SULT2B1) that share many common structural features bolic system. (reviewed in17). The properties of the various human SULTs are summarised in Table 1. The results of many studies show THE SULFOTRANSFERASE SUPERFAMILY— clearly that the expression of SULTs in humans is carefully ORGANISATION AND FUNCTION regulated with respect to tissue type, development and As with many drug metabolising enzymes, the cytosolic hormonal influences, supporting proposed functional roles SULTs are derived from a large superfamily of genes. Full- for some of the enzymes. A number of examples are worth length cDNAs encoding more than 50 mammalian and avian cytosolic SULTs have been cloned and sequenced, and *Raftogianis RB et al, submitted for publication. In particular, the many of the expressed proteins characterised. The majority new nomenclature for the SULT1C enzymes may cause some of these enzymes sulfate small molecule xenobiotics and/or confusion. The human SULT1C2 referred to here was called endogenous hormones and neurotransmitters, and based on SULT1C1 or SULT1C sulfotransferase 1 in the original descrip- amino acid sequence analysis can be subdivided into six tions,18,19,75 and the human SULT1C4 referred to here was originally families (Box 2). In addition plants possess a number of called SULT1C2 or SULT1C sulfotransferase.19,75

The Pharmacogenomics Journal Recent advances in sulfotransferase research MWH Coughtrie 299

considering further: SULTs 1C2 and 1C4 appear to be most 17b-estradiol, suggesting an important role for this enzyme highly expressed in fetal tissues,13 although RNA dot blots in modulating estrogen action.21 Indeed, the enzyme is indicated the adult stomach and kidney (SULT1C2) and expressed in the endometrium and is exquisitely regulated ovary (SULT1C4) may be sites of expression.18,19 However during the menstrual cycle22,23 (probably under the primary the function of these enzymes in humans is not clear. The influence of progesterone24,25) where it is postulated to catecholamine sulfotransferase SULT1A3 is expressed at moderate estrogenic stimulation of the endometrium high levels in fetal liver but hepatic expression is essentially around the time of implantation.23 A further insight into absent in the adult—here the gastrointestinal tract is the the importance of estrogen sulfation has come from the major site12 which correlates with the dopaminergic func- identification that a number of hydroxylated metabolites of tion of the gut, where the majority of dopamine sulfate is the ubiquitous environmental pollutants polychlorinated 4 produced. The fetal adrenal gland produces large amounts biphenyls are extraordinarily potent inhibitors (Ki in the pM of DHEA sulfate to support placental estrogen biosynthesis, range) of SULT1E1,26 suggesting a mechanism whereby a requirement reflected in the high level expression of the these chemicals exert their well-known endocrine disrupt- SULT2A1 isoform in the fetal zone of 2nd and 3rd trimester ing effects.27 adrenal.9,20 Another important example involves the estro- A particularly intriguing problem currently occupying a gen sulfotransferase (SULT1E1) which displays a particularly number of investigators in the field is to determine the high affinity (in the low nM range) for its natural substrate function of the most recently discovered SULT family, the

www.nature.com/tpj Recent advances in sulfotransferase research MWH Coughtrie 300

Table 1 Properties of the human cytosolic sulfotransferases

SULT isoform Chromosomal location Diagnostic substratea Major sites of expressionb

SULT1A1 16p12.1–11.2 4-Nitrophenol Adult liver, adult GI tract, adult platelets, placenta SULT1A2 16p12.1–11.2 * ?c SULT1A3 16p11.2 Dopamine Adult GI tract, adult platelets, adult brain, placenta, fetal liver SULT1B1 4q11–13 * Adult liver, adult GI tract, fetal GI tract SULT1C2 2q11.2 * Fetal kidney, fetal lung, fetal GI tract SULT1C4 2q11.2 * Fetal kidney, fetal lung SULT1E1 4q13 17b-Estradiol Fetal liver, fetal lung, fetal kidney, adult liver, endometrium SULT2A1 19q13.3 Dehydroepiandrosterone Fetal adrenal, fetal liver, adult liver, adult adrenal SULT2B1d 19q13.3 Cholesterol Adult skin, prostate, placenta SULT4A1 22q13.1–13.2 No substrate known Brain aIndicates a substrate that may be used to assess enzyme activity of the isoform in human tissue cystols. Where an asterisk (*) appears in this column, no substrate has yet been identified that is sufficiently specific for the isoform concerned to be of value in quantifying the associated enzyme activity in tissue samples. 4-Nitrophenol may be used to follow activity of these expressed recombinant proteins in vitro. bExpression is taken to mean enzyme protein expression. Where RNA expression only has been demonstrated (eg for SULTs 1C2 and 1C4 in certain adult tissues18.19 ), this has not been included. cThere is considerable doubt as to whether the SULT1A2 protein is actually expressed at significant levels in human tissues.71 d It appears that the splice variant SULT2B1b is the major form expressed at the protein level.93,94 The enzyme also sulfates DHEA, although cholesterol may be diagnostic.95

4A proteins. SULT4A1 cDNAs have been isolated from three important chemical defence function, particularly in the mammalian species so far—rat, mouse and human—and the absence from the liver of the other major conjugating predicted proteins share a remarkable degree of similarity. enzymes such as UDP-glucuronosyltransferases15 (although The mouse and the rat proteins have identical amino acid the fetal kidney does express some UDP-glucuronosyltrans- sequences, and the human has only six (mainly conserva- ferase29). These studies are extremely difficult to carry out, tive) amino acid changes out of 284. Human SULT4A1 however the wide distribution and tight regulation of SULT shares no more than 34% amino acid sequence identity with expression during human development points to a particu- any other member of the family. The SULT4A1 proteins larly important role for these enzymes in the early stages of appear to be expressed only in the brain,28 and despite life. significant effort no natural or xenobiotic substrate (ligand) has yet been identified. Together, these observations strongly suggest a specific and important function for this STRUCTURE–FUNCTION STUDIES ON protein, which may require the application of knockout SULFOTRANSFERASES and/or transgenic technologies to fully understand. Structural biology is one of the most powerful tools available The concept of sulfation as a particularly important for helping to understand macromolecular function, and metabolic reaction in the developing human fetus is one structural approaches have recently revealed some interest- that has also matured over recent years. Many SULTs are ing and potentially valuable secrets of the SULTs. The first expressed at high levels in the fetus (Table 1), and in fact SULT structure to be solved, by the group of Pedersen and some appear to be expressed only or primarily in the Negishi at NIEHS using X-ray crystallography, was that of prenatal period, such as the SULT1C enzymes.13,18,19 The mouse SULT1E1.30 This enzyme is unusual in the SULT function of certain SULTs in fetal development is fairly well world in that, in vitro, it exists as a monomer in solution— established, such as the production of DHEA sulfate by the most of the enzymes appear to form homodimers. Structures adrenal SULT2A1 enzyme.9 However, the role of others is of two other human cytosolic SULTs have since been solved: not so clear. SULT1C enzymes are expressed (in sexually SULT1A331,32 and SULT2A1,33 as well as the sulfotransferase dimorphic manner) in adult rodent liver where they seem to domain of human heparan sulfate N-deacetylase/N- be involved in xenobiotic metabolism—particularly the sulfotransferase34 and the SULT-like insect enzyme retinol bioactivation of procarcinogens. This may not be the case dehydratase.35 Figure 1 shows the overall SULT protein fold, in humans, where the enzymes are reputedly involved in as demonstrated by the mouse SULT1E1 crystal structure. thyroid hormone metabolism—this is a particularly impor- The three cytosolic SULT structures are highly conserved, tant function in the fetus where appropriate thyroid and a number of key features of the SULT active site and hormone homeostasis is essential for normal brain develop- reaction mechanism (reviewed in36) can be deduced from ment. The human fetus produces very large amounts these and other studies (see Figure 2). The amino acids (compared with the adult) of iodothyronine sulfates (our involved in binding the sulfuryl donor PAPS are, as would be unpublished work in collaboration with R Hume). The high expected, highly conserved throughout the SULT family. levels of SULT1A1 in various fetal tissues12 may provide an The PAPS 50-phosphate binding site (consensus sequence

The Pharmacogenomics Journal Recent advances in sulfotransferase research MWH Coughtrie 301

Lys258

Arg257

α-6 Helix

PAP Lys48

Thr52

Tyr Glu PSB-loop 240 146 Dopamine His108

Lys106 Phe24 Phe81

Figure 1 Structure of mouse SULT1E1. The co-ordinates for sulfotransferase 1E1 from mouse crystallised in the presence of Figure 2 Active site model of human SULT1A3. To construct the active site model, the co-ordinates of mouse SULT1E1 (PDB entry PAP and substrate 17b-estradiol were obtained from the Protein 31 Data Bank (http://www.rcsb.org/pdb/, entry 1AQU). A representa- 1AQU) and the partial human SULT1A3 structure were used. tion of the structure of a SULT1E1 monomer is shown, with a-helical Molecule A of 1AQU was used as a template, and the SULT1A3 a sections displayed as red cylinders and b-sheets as light blue arrows. structure was super-imposed as the -carbons of residues 42–45, The figure was created using WebLab Viewer Lite (Accelrys, San 45–50, 52–58, 104–107, 136–147 and 166–170. Side chains of the b Diego, USA). SULT1E1 component covering PAP and 17 -estradiol (residues 225–264) were mutated to their corresponding residues in SULT1A3, and dopamine was docked into the active site based on the position of estradiol in 1AQU, so that the 3-OH is on the estradiol phenolic OH and the 4-OH is towards Tyr240. The figure TYPKSGTTW, corresponding to residues 45–53 in SULT1A3) was created using WebLab Viewer Lite (Accelrys, San Diego, USA). is very similar to the P-loop motif found in many nucleotide binding proteins, and is called the PSB-loop in SULTs.30 The lysine residue within this motif (Lys48 in human SULT1A3 and mouse SULT1E1) is absolutely conserved and is believed challenging problem to determine what controls the sub- to act as a catalytic acid in the reaction, by protonating the strate specificity of these various enzymes, but this is an phosphate–sulfate bridge oxygen of PAPS thereby enhancing important goal from a number of perspectives, including dissociation of the leaving group.37 Residues from various being able to predict the metabolic fate of drugs metabolised parts of the polypeptide chain form the binding site for the by sulfation, or even to assist the design of selective 0 3 -phosphate of PAPS. Arg130 and Ser138 (Ser138 is absolutely inhibitors. The protein folds of the SULTs whose structures conserved in sulfotransferases) are particularly important, have been solved are essentially superposable, however along with the guanidinium group of Arg257 and the amide some subtle, but important differences in the active sites nitrogens of Lys258 and Gly259 that contribute to a motif of the enzymes do exist. The human SULT1A3 enzyme highly conserved within the cytosolic SULTs (RKGxxGxWK, provides a fine example: it displays a high selectivity for residues 257–265 in SULT1A3). The catalytic centre of the endogenous and xenobiotic catecholamines (particularly enzyme is a histidine residue (His108) that acts as a catalytic dopamine) not shared by its close relative SULT1A1 even base, abstracting a proton from the hydroxyl group on the though the two enzymes are more than 93% identical at the acceptor substrate leaving a nucleophilic oxygen that amino acid sequence level—ie only 20 residue differences ‘attacks’ the sulfur on PAPS. The presence of a histidine out of 295. Analysis of these sequence differences38 showed residue in this position is an absolutely conserved feature of that a number of them occur in a short stretch (143– all SULTs, and mutating it abolishes enzyme activity.37 The HRMEKAHPEP–152) immediately C-terminal to a conserved reaction intermediate (at least in mouse SULT1E137)is sequence that we now know to form the a-6 helix.31 stabilised by His108, Lys106 and Lys48. Mutation of two amino acids in this stretch of SULT1A3 to Of course, a vital feature of the SULT enzyme family is the their corresponding residues in SULT1A1 (His143-Tyr and degree of substrate specificity demonstrated by the indivi- Glu146-Ala) followed by detailed kinetic analysis of the dual isoforms. For example, humans have enzymes that are purified, recombinant proteins clearly demonstrated that highly selective for catecholamines (SULT1A3), estrogens Glu146 was the major determinant of the specificity of (SULT1E1) and DHEA (SULT2A1). Conversely the SULT1A1 SULT1A3 for dopamine38 (Figure 2). The group of McManus enzyme appears to be much more promiscuous, able to performed similar studies, including the reverse experiment sulfate a wide variety of xenobiotics with high affinity. It is a on SULT1A1, arriving at the same conclusion.39,40 Subse-

www.nature.com/tpj Recent advances in sulfotransferase research MWH Coughtrie 302

quent molecular modelling experiments based on the X-ray SULT1A1 activity. Development of specific antibodies crystal structure of SULT1A3 suggested a molecular basis for recognising human SULTs on immunoblot analysis allowed this specificity, as the basic ethylamine function in dopa- confirmation that, in most tissues at least, the variability in mine and similar molecules would be in close proximity to SULT1A1 enzyme activity strongly correlates with the level w.50,51 the acidic carboxyl side chain of Glu146, and able to form a of cytosolic enzyme protein The next major advances 31 salt bridge. The influence of Glu146 goes further however, came with the cloning of human SULT1A1 cDNAs, which and probably controls the orientation of many compounds provided the first evidence that coding region sequence in the SULT1A3 active site, explaining the selectivity of variant alleles of SULT1A1 existed in the population. There SULT1A3 for catechols over corresponding phenols.31 The are now eight characterised SULT1A1 cDNA sequences in the only other major study on SULT selectivity suggests that the GenBank/EMBL sequence databases, which mainly represent specificity of mouse SULT1E1 for 17b–estradiol over DHEA the two most common SULT1A1 alleles determined by gene arises from a ‘gating’ phenomenon. Petrotchenko et al41 sequencing studies, termed SULT1A1*1 and SULT1A1*2 demonstrated that Tyr81 and Phe142 form a narrow gate that (Table 2). Some of these cDNA sequences contain nucleotide ‘allows’ 17b-estradiol entry to the active site but ‘prevents’ variants that have never been found in gene sequencing

DHEA (which has a 19-methyl group) from entering. Phe142 studies, which either represent extremely rare SULT1A1 is absolutely conserved in SULTs, and Tyr81 alone was found alleles, or (perhaps more likely) have arisen from PCR and/ to regulate the gate phenomenon. or sequencing errors (GenBank/EMBL accession numbers Clearly, our understanding of the molecular basis for SULT L19999, U26309 and X84654). Two entries encode the wild- substrate specificity is still in its infancy. The structural tools type SULT1A1*1 allozyme (X78283 and AJ007418) and three now available to us will certainly allow major advances to be encode the common variant allozyme SULT1A1*2 (L10819, made in this important area of sulfation biology in the L19955, U09031). coming years. Seminal studies from Weinshilboum’s group confirmed that platelet phenol sulfotransferase enzyme activity and INTER-INDIVIDUAL VARIATION IN thermal stability were related to SULT1A1 genotype. They * SULFOTRANSFERASES—MECHANISMS AND showed that individuals homozygous for the SULT1A1 2 CONSEQUENCES allele had only about 15% of the platelet phenol sulfo- * * * Ever since studies of human SULT biochemistry began it has transferase activity of 1A1 1/1A1 2 heterozygotes or 1A1 1 been obvious that the activity of these enzymes varies homozygotes, and also that the thermal stability of enzyme 52,53 widely in the population. Early work focused on the phenol activity was associated with SULT1A1 genotype. A SULTs (those we now call SULTs 1A1 and 1A3), and was subsequent study confirmed that reduced platelet SULT * facilitated by the use of platelets as enzyme source. These activity is observed in 1A1 2 homozygotes,54 although the studies have been thoroughly reviewed elsewhere.42–44 The effect was not as pronounced with only a 50% reduction in key questions relating to inter-individual variability in any activity apparent. It remains to be conclusively proven how SULT are, of course: (a) what is the molecular basis?; and the single amino acid change (Arg213His) present in (b) what are the physiological consequences? We are SULT1A1*2 results in decreased enzyme activity, protein considerably further advanced in our ability to answer the expression and thermal stability. Experiments with the former, although several recent molecular epidemiological recombinant proteins have, in the main, failed to show studies have begun to suggest roles for SULT polymorphisms significant differences in kinetic properties between SULTs * * in disease susceptibility. 1A1 1 and 1A1 2, although the thermal stability of the SULT1A1*2 allozyme appears considerably reduced com- SULT1A1 Pharmacogenetics pared to the wild-type enzyme.53,55,56 Recent work from this SULT1A1 is the most widely studied polymorphic SULT, and laboratory using recombinant enzymes expressed and is significantly due to its broad substrate specificity, wide purified from E. coli suggests that the 1A1*2 allozyme may tissue distribution and involvement in the metabolism and be a more catalytically efficient enzyme, and it is certainly detoxification of many drugs and other xenobiotics as well considerably less susceptible to the partial substrate inhibi- as in the bioactivation of dietary and environmental tion phenomenon displayed by SULT1A1*1 (Tabrett CA and procarcinogens.45 Biochemical and genetic studies (using Coughtrie MWH, unpublished work). However, the 1A1*2 twins and other family cohorts) indicated that variation in allozyme protein appears to have a shorter biological half- this ‘thermostable’ phenol SULT activity was regulated by life than the wild-type, and preliminary data suggest that it inheritance,46 and also suggested that this variation might is more readily degraded via the proteasome pathway.57 This be associated with sulfation capacity for drugs such as would explain the various observations—reduced activity, paracetamol in vivo.47,48 In addition to variation in SULT1A1 reduced protein expression and thermal instability—that enzyme activity, it was also discovered that the thermal have been reported. Examination of the SULT crystal stability of this enzyme activity was under the control of a genetic polymorphism.49 Differences in thermal stability wThis may not be the case for liver, where the presence of other SULT can arise from subtle changes in protein structure (eg single isoforms such as 1B1 may interfere with measurement of enzyme amino acid changes), and these studies gave the first clues as activity with non-selective substrates such as 4-nitrophenol (our to the molecular basis for inter-individual variation in unpublished work).

The Pharmacogenomics Journal Recent advances in sulfotransferase research MWH Coughtrie 303

Table 2 Human sulfotransferase allozymes colorectal cancer patients. These results support data indicating that SULT1A1*1 homozygosity is associated with SULT Allozyme Amino acid change(s) Allozyme frequencya reduced risk of colorectal cancer.62 However, a large study of isoform breast cancer patients showed no overall effect of SULT1A1 genotype on risk of developing the disease, although these SULT1A1 *1 – 0.674 authors did demonstrate that genotype may be associated *2 Arg213His 0.313 with age of onset of breast cancer.63 In contrast, investiga- *3 Met Val 0.010 223 tion of another breast cancer patient cohort showed *4 Arg37Gln 0.003 * SULT1A2 *1 – 0.508 increased risk associated with SULT1A1 1 genotype when combined with intake of well-done meat, a significant *2 Ile7Thr; Asn235Thr 0.287 64 *3 Pro19Leu 0.180 source of many heterocyclic amines that are bioactivated 45,65 *4 Ile7Thr; Arg184Cys; Asn235Thr 0.008 by human SULTs including SULT1A1. To complicate 64 *5 Ile7Thr 0.008 matters, however, this same study found that overall risk * *6 Asn235Thr 0.008 of breast cancer was associated with the SULT1A1 2 allele. A SULT1C2 *1 – 0.905 small study of prostate cancer failed to find any association *2 Ser255Ala 0.067 with SULT1A1 genotype.66 Clearly, these diseases are multi- *3 Asp60Ala 0.011 factorial and it would be surprising to find very strong *4 Arg Gln 0.011 73 associations with SULT1A1 genotype. Only through large- *5 Ser111Phe 0.006 SULT2A1 *1 – 0.935 scale studies of diseases where sulfation may be important, b and accounting for specific environmental exposures, will *2 Met57Thr 0.027 b *3 Glu186Val 0.038 we gain greater insight into the function of SULTs and c – Ala63Pro 0.050 sulfation in humans. c 67,68 – Lys227Glu 0.008 In common with other drug metabolising enzymes, c – Ala261Thr 0.133 there appear to be inter-ethnic/inter-racial differences in the incidence of the various SULT1A1 alleles (Table 3). African– aAllozyme frequencies in Caucasian populations (except where stated) from published studies. Taken from53 (SULT1A1, 1A2)74, (SULT1C2) and81,83 (SULT2A1). Americans appear to have a particularly high frequency of * * bSULT2A1 allozyme assignments of *2 and *3 are preliminary, as first reported by the 1A1 3 allele, associated with reduced 1A1 1 allele Nagata and Yamazoe.82 frequency—1A1*2 allele frequency is not significantly cThese SULT2A1 variants have so far only been found in African–American different to Caucasian populations. In contrast, the 1A1*1 subjects,83 and have not yet been allocated an allozyme number. allele is present at significantly higher frequencies in oriental populations (Chinese, Japanese) compared with structure does not provide an obvious mechanistic explana- either the Caucasian or African–American populations tion; amino acid 213 is on the surface of the molecule studies so far. The significance of these differences in terms distant from the active site and from the proposed of sulfation function is not yet clear, and further studies in dimerisation domain.58 this area are certainly required. However, these observations The fact that SULT1A1 is: (a) important for xenobiotic suggest that different ethnic/racial populations may display detoxification, thyroid hormone metabolism59 and the different capacities for sulfation via SULT1A1. bioactivation of procarcinogens,45 and (b) subject to a common functional polymorphism has inspired a number of investigators to carry out molecular epidemiological SEQUENCE AND FUNCTIONAL VARIATION IN OTHER studies in an attempt to associate SULT1A1 genotype with HUMAN SULFOTRANSFERASES various allegedly chemical-associated pathologies, including SULT Family 1 colorectal, breast and prostate cancer. Predictably, such Coding region single nucleotide polymorphisms resulting in studies have produced conflicting results. It is also interest- amino acid changes (ie non-synonymous cSNPs) have been ing that a significant age-related effect on SULT1A1 identified in a number of Family 1 SULTs. Five cSNP variants genotype has been observed, whereby the frequency of the of SULT1A2 are known, from either cDNA or gene sequen- SULT1A1*1 allele was higher in the older age-groups.60 This cing studies,53,69,70 (Table 2), and experiments in recombi- observation raises the possibility that the genotype respon- nant systems suggest that the protein produced by the * sible for a high sulfation activity phenotype may provide common variant SULT1A1 2 (Ile7Thr, Asn235Thr) has a 53,69 some protection against long-term cell/tissue damage from substantially higher Km for the substrate 4-nitrophenol. xenobiotic and/or endogenous chemicals, and as such The functional significance of this is not clear, and would support a major role for SULT1A1 in detoxification. investigating it is complicated: (a) by the possibility that a Although this finding requires to be confirmed in a larger significant portion of SULT1A2 RNA is incorrectly spliced and more tightly defined ageing population, it does and therefore protein production is hampered71 (indeed we illustrate the importance of careful age matching for any have seen no evidence for SULT1A2 protein by western blot case-control studies involving genotyping. analysis in various human tissues13); and (b) by the Frame et al61 reported that low platelet phenol SULT observation that the most common variant alleles of activity (attributed to SULT1A1) was more prevalent in SULT1A1 and SULT1A2 (ie the *2 alleles) are in strong

www.nature.com/tpj Recent advances in sulfotransferase research MWH Coughtrie 304

Table 3 Ethnic differences in SULT1A1 allele frequencies frame.htm) indicate the presence of very rare (allele frequency 0.008) SULT1E1 cSNPs in the Caucasian (Ala32Val Study population 1A1*1 1A1*2 1A1*3 Chromosomes Reference and Pro253His) and African–American (Asp22Tyr) popula- tested tions. Again, no functional characterisation of the resulting variant allozymes has been reported. Caucasian 0.656 0.332 0.012 490 72 Caucasian 0.674 0.313 0.010 300 53 SULT Families 2 and 4 a Caucasian 0.679 0.321 – 586 60 Circulating DHEA sulfate levels vary substantially between a Caucasian 0.630 0.370 – 600 70 individuals, and it is suggested that this variation is, at least Caucasian 0.654 0.317 0.029 104 96 in part, inherited.79 It is reasonable to hypothesise that African–American 0.477 0.294 0.229 140 72 Nigeriana 0.731 0.269 – 104 60 inter-individual variation in the expression and/or activity Chinese 0.914 0.080 0.006 580 72 of SULT2A1 might therefore contribute to these differences Japanese 0.832 0.168 0.000 286 55 in DHEA sulfate levels. Indeed, DHEA SULT activity in human liver cytosols appears to follow a bimodal distribu- aStudies performed using PCR-RFLP assays unable to discriminate between the *1 tion, and a strong correlation between enzyme activity and and *3 alleles. Data from these studies in 1A1*1 column will represent combined protein expression exists.80 Analysis of SULT2A1 cDNA allele frequencies for *1 and *3. sequences isolated by a number of laboratories shows the presence of the presumed wild-type sequence (SULT2A1*1, positive linkage disequilibrium.53,70,72 No cSNP polymorph- GenBank/EMBL accession numbers U08024, U08025 and isms in SULT1A3 have been published to date,73 and in fact X84816—consistent with the two SULT2A1 gene sequences this gene seems remarkably devoid of sequence variation U13056–U13061 and L36191–L36196) and two apparent between individuals. cSNPs (Thr90Ser, L20000 and X70222; Leu159Val, S43859, 81 The SULT1B1 genes from 48 Japanese have been L02337). However, extensive gene sequencing studies sequenced and a single cSNP identified (Glu204Asp), failed to identify the variant alleles represented by these 73 although no allele frequency for this variant was reported. cDNAs. These authors did report two cSNPs (Met57Thr, A number of non-coding region SNPs were also identified in Glu186Val) with allele frequencies around 3% which have * * this study. been assigned the preliminary designation 2 and 3, 82 * * Sequencing of the human SULT1C2 gene from different respectively, (Table 2). The SULT2A1 2 and 3 allozymes individuals has revealed four cSNPs (Table 2), and recombi- showed decreased enzyme activity when expressed in COS 81 nant allozymes representing three of these displayed, cells. A further three cSNPs in SULT2A1 (Ala63Pro, reduced enzyme activity compared to the wild-type allo- Lys227Glu and Ala261Thr) were originally reported by the zyme.74 A common 30bp insertion/deletion variant within Pharmacogenetics Knowledge Base (http://www.pharmgkb. intron three of SULT1C2 also exists,74 and there is a rare org/data/mayo/SULT2A1a_files/frame.htm) and have recently 83 SULT1C2 transcript variant where an additional exon (exon been published (Table 2). They have only been found in 75 3b) is spliced into the mRNA molecule. The consequences African–American populations; Ala261Thr is common, with for the individual of possessing these variant alleles and/or an allele frequency of 13%. This may be significant as inter- 80 expressing the transcript variant are unknown, although the ethnic differences in SULT2A1 activity were reported. SULT1C2 enzyme is expressed at high levels in fetal tissues Interestingly, the most common of the variants found in and has been implicated in the bioactivation of aromatic African–American subjects (Ala261Thr) influences the for- 83 amines,19,76 and also in the metabolism of iodothyro- mation of SULT2A1 dimers in vitro, as it occurs within the 59,77 58 nines. A single cSNP in SULT1C4 (Asp5Glu) was conserved dimerisation motif. No non-synonymous cSNPs identified in a sequencing study of many SULT genes in 48 in SULT2A1 or SULT2B1 were found in the Japanese 73 Japanese individuals, although no allele frequency or cohort. allozyme function data were reported.73 In one study, a large number of SNPs were found in the The expression and activity of SULT1E1 varies widely in SULT4A1 gene (whose function remains to be determined), 73 the human population,23,78 although it is not known although all 42 occurred in non-coding regions. This is whether this is under genetic control. No evidence for perhaps surprising, given the extremely highly conserved gender-specific regulation in liver was found,78 although the nature of the SULT4A1 protein from mouse to human. enzyme is expressed in female-specific tissues, including the Another SULT with an important endogenous function, endometrium. It is possible that the variability in SULT1E1 SULT1A3, shows an extremely low level of variation and one expression results from different chemical influences, since might have expected the same pattern for SULT4A1. progesterone, other hormones and alcohol are known to influence expression levels in vitro and/or in vivo.24,25,78 The FUTURE DIRECTIONS IN SULFOTRANSFERASE only published study on SULT1E1 nucleotide sequence RESEARCH variants reported a number of non-coding region variants Significant advances have been made over recent years in (50- and 30-flanking regions, introns) but no cSNPs.73 our understanding of the SULT enzyme family, and of the However, data reported by the Pharmacogenetics Knowledge sulfation system as a whole, including (but not restricted to) Base (http://www.pharmgkb.org/data/mayo/SULT1E1a_files/ those outlined here. There is no reason to doubt this trend

The Pharmacogenomics Journal Recent advances in sulfotransferase research MWH Coughtrie 305

will continue, and it may be driven, at least in part, by the abnormal sperm function (reduced motility) and conse- needs of the Pharmaceutical Industry. The fact that many quently the older male mice produced smaller litters. In adverse drug reactions are directly related to genetic and/or young SULT1E1 knockout mice, however, a reduction in environmental effects on the major cytochromes P450 fertility was also observed although this was associated with (particularly CYP2D6 and the CYP3A family)84 has led many the female animals. These elegant studies certainly provide in the industry, as well as the regulatory authorities, to solid evidence that estrogen sulfation is important for (male) re-evaluate drug discovery and development priorities away reproductive function in mice. However, as with all animal from compounds whose primary route of metabolism is via experiments, the relevance to humans must be carefully the CYP system.85 The consequence of this is that many assessed and clearly demonstrated. Although there is some drugs coming to market in the future will be metabolised evidence (RT-PCR data only) that SULT1E1 is expressed in primarily by other pathways, in particular the conjugating the human testis91 there are fundamental differences enzymes such as UDP-glucuronosyltransferases and SULTs. between mice and humans in the expression profile of this This is already beginning to happen and the example of the important enzyme. For example, hepatic expression is anti-diabetic drug troglitazone illustrates the point. Inhibi- sexually dimorphic in mice92 but not in humans,78 in tion of bile salt transport by troglitazone sulfate has been humans the enzyme is expressed at much higher levels in implicated in the severe idiosyncratic hepatoxicity of this fetal than adult tissues,13 and a major and critical site of drug86 which led to its withdrawal from the US market in expression in adult humans is the endometrium.23 It is 2000, only 3 years after its launch. It will therefore be certain this will be the first of many knockout/transgenic necessary to know much more about the biology and experiments involving SULTs, and much useful information genetics of these enzymes in both human and experimental will undoubtedly come from such studies. animal species if we are to be able to predict adverse So, there is certainly much of interest in the SULT field reactions involving compounds metabolised by conjuga- currently, as there will be in the foreseeable future, to appeal tion. This will require the application of a range of to the curious scientist and/or clinician. The highlights experimental approaches, including functional genomics, presented here are, of necessity, a personal selection, nucleic acid and protein array technologies, high- although hopefully they cover the more pertinent aspects resolution/high-sensitivity analytical techniques such as of this fascinating system relating to human biology. These mass spectrometry and NMR spectroscopy, structural biol- and other studies directed towards understanding the ogy, in vivo metabolic studies and transgenic/knockout function of sulfation and the SULTs will undoubtedly reveal experiments. Studies of in vivo sulfation in humans are thin important roles for the enzymes that are not yet obvious, on the ground—lack of good, safe, specific probe substrates but that we will look back on from a more enlightened is one major stumbling block—but such analyses are of key future. importance in assessing the functional consequences of inter-individual variation in SULTs. Much useful informa- ACKNOWLEDGEMENTS tion on the functions of cytochromes P450 have come from I am grateful to all the members of my laboratory (past and present), such studies, and there is a desperate need for them to be and to numerous valued collaborators, who have contributed to the conducted on sulfation. Hopefully this will be an area where work and ideas described here. I am particularly indebted to Dr Rebecca significant advances can be made in the future. Raftogianis (Fox Chase Cancer Center, Philadelphia, USA) for her critical The example of troglitazone also highlights another im- reading of the manuscript and for helpful suggestions regarding nomenclature, and to Professor Jyrki Taskinen (University of Helsinki, portant area for future research—the transport of substrates and Finland) for producing the model upon which Figure 2 is based. Current products of the sulfation reaction into and out of cells. There research in the laboratory is supported by the Medical Research Council, has been an explosion in research into small molecule transport the Commission of the European Communities (QLG3-CT-2000- 87–89 in recent years (see reviews ) and the major families of 00930), Tenovus Scotland, the Scottish Office Chief Scientist Office transport proteins (MRPs, OATPs and BSEP) are important and the Human Drug Conjugation Consortium. components of the sulfation system. It will be important to understand the substrate specificities of these multiple trans- DUALITY OF INTEREST None declared. porters for sulfate conjugates and SULT substrates, and also to explore the interactions of endogenous transporter substrates such as bile salts, steroid and thyroid hormones ABBREVIATIONS with sulfate conjugates generated during xenobiotic metabo- SULT Sulfotransferase lism. PAPS 30-phosphoadenosine 50-phosphosulfate Finally, it is worth highlighting one very recent landmark advance in the field: the generation of the first SULT knockout mice.90 Using standard targeting vector metho- dology, these investigators generated mice deficient in REFERENCES estrogen SULT (SULT1E1). The main phenotype of the 1 Robbins PW, Lipmann F. Isolation and identification of active sulfate. J Biol Chem 1957; 229: 837–851. knockout mice was observed in males, where age-dependent 2 Klaassen CD, Boles JW. The importance of 30-phosphoadenosine 50- Leydig cell hypertrophy/hyperplasia coupled with semini- phosphosulfate (PAPS) in the regulation of sulfation. FASEB J 1997; 11: ferous, tubule damage occurred. These lesions resulted in 404–418.

www.nature.com/tpj Recent advances in sulfotransferase research MWH Coughtrie 306

3 Dousa MK, Tyce GM. Free and conjugated plasma catecholamines, 23 Rubin GL, Harrold AJ, Mills JA, Falany CN, Coughtrie MWH. Regulation DOPA and 3-O-methyldopa in humans and various animal species. Proc of sulfotransferase expression in the endometrium during the menstrual Soc Exp Biol Med 1988; 188: 427–434. cycle, by oral contraceptives and during early pregnancy. Mol Human 4 Eisenhofer G, Coughtrie MWH, Goldstein DS. Dopamine sulphate: an Reprod 1999; 5: 995–1002. enigma resolved. Clin Exp Pharmacol Physiol 1999; 26: S41-S53. 24 Tseng L, Liu HC. Stimulation of arylsulfotransferase activity by 5 Ogura K, Kajita J, Narihata H, Watabe T, Ozawa S, Nagata K et al. progestins in human endometrium in vitro. J Clin Endocrinol Metab Cloning and sequence analysis of a rat liver cDNA encoding 1981; 53: 418–421. hydroxysteroid sulfotranferase. Bio chem Biophys Res Commun 1989; 25 Falany JL, Falany CN. Regulation of estrogen sulfotransferase in human 165: 168–174. endometrial adenocarcinoma cells by progesterone. Endocrinology 6 Ogura K, Kajita J, Narihata H, Watabe T, Ozawa S, Nagata K et al. cDNA 1996; 137: 1395–1401. cloning of the hydroxysteroid sulfotransferase STa sharing a strong 26 Kester MHA, Bulduk S, Tibboel D, Meinl W, Glatt H, Falany CN et al. homology in amino acid sequence with the senescence marker protein Potent inhibition of estrogen sulfotransferase by hydroxylated PCB SMP-2 in rat livers. Biochem Biophys Res Commun 1990; 166: 1494– metabolites: a novel pathway explaining the estrogenic activity of PCBs. 1500. Endocrinology 2000; 141: 1897–1900. 7 Watabe T, Ogura K, Satsukawa M, Okuda H, Hiratsuka A. Molecular 27 Safe SH. Polychlorinated biphenyls (PCBs): environmental impact, cloning and functions of rat liver hydroxysteroid sulfotransferases biochemical and toxic responses, and implications for risk assessment. catalysing covalent binding of carcinogenic polycyclic arylmethanols Crit Rev Toxicol 1994; 24: 87–149. to DNA. Chem Biol Interact 1994; 92: 87–105. 28 Falany CN, Xie XW, Wang J, Ferrer J, Falany JL. Molecular cloning and 8 Otterness DM, Wieben ED, Wood TC, Watson RWG, Madden BJ, expression of novel sulphotransferase-like cDNAs from human and rat McCormick DJ et al. Human liver dehydroepiandrosterone sulfotransfer- brain. Biochem J 2000; 346: 857–864. ase: molecular cloning and expression of cDNA. Mol Pharmacol 1992; 29 Hume R, Coughtrie MWH, Burchell B. Differential localization of UDP- 41: 865–872. glucuronosyltransferase in kidney during human embryonic and fetal 9 Barker EV, Hume R, Hallas A, Coughtrie MWH. Dehydroepiandrosterone development. Arch Toxicol 1995; 69: 242–247. sulfotransferase in the developing human fetus—quantitative biochem- 30 Kakuta Y, Pedersen LG, Carter CW, Negishi M, Pedersen LC. Crystal ical and immunological characterization of the hepatic, renal, and structure of estrogen sulphotransferase. Nature Struct Biol 1997; 4: adrenal enzymes. Endocrinology 1994; 134: 982–989. 904–908. 10 Hume R, Barker EV, Coughtrie MWH. Differential expression and 31 Dajani R, Cleasby A, Neu M, Wonacott AJ, Jhoti H, Hood AM et al. X-ray immunohistochemical localization of the phenol and hydroxysteroid crystal structure of human dopamine sulfotransferase, SULT1A3: sulfotransferase enzyme families in the developing lung. Histochem Cell molecular modelling and QSAR analysis demonstrate a molecular basis Biol 1996; 105: 147–152. for sulfotransferase substrate specificity. J Biol Chem 1999; 274: 37862– 11 Hume R, Coughtrie MWH. Phenolsulfotransferase—localization in 37868. kidney during human embryonic and fetal development. Histochem J 32 Bidwell LM, McManus ME, Gaedigk A, Kakuta Y, Negishi M, Pedersen L 1994; 26: 850–855. et al. Crystal structure of human catecholamine sulfotransferase. JMol 12 Richard K, Hume R, Kaptein E, Stanley EL, Visser TJ, Coughtrie Biol 1999; 293: 521–530. MWH. Sulfation of thyroid hormone and dopamine during human 33 Pedersen LC, Petrotchenko EV, Negishi M. Crystal structure of development—ontogeny of phenol sulfotransferases and arylsulfatase SULT2A3, human hydroxysteroid sulfotransferase. FEBS Lett 2000; in liver, lung and brain. J Clin Endocrinol Metab 2001; 86: 475: 61–64. 2734–2742. 34 Kakuta Y, Sueyoshi T, Negishi M, Pedersen LC. Crystal structure of the 13 Stanley EL. Expression and function of iodothyronine metabolising sulfotransferase domain of human heparan sulfate N-deacetylase/N- enzymes during human placental and fetal development. PhD Thesis, sulfotransferase 1. J Biol Chem 1999; 274: 10673–10676. University of Dundee, 2001. 35 Pakhomova S, Kobayashi M, Buck J, Newcomer ME. A helical lid 14 Hakkola J, Pelkonen O, Pasanen M, Raunio H. Xenobiotic-metabolizing converts a sulfotransferase to a dehydratase. Nat Struct Biol 2001; 8: cytochrome P450 enzymes in the human feto-placental unit: role in 447–451. intrauterine toxicity. Crit Rev Toxicol 1998; 28 : 35–72. 36 Yoshinari K, Petrotchenko EV, Pedersen LC, Negishi M. Crystal 15 Coughtrie MWH, Burchell B, Leakey JEA, Hume R. The inadequacy of structure-based studies of cytosolic sulfotransferase. J Biochem Molec perinatal glucuronidation—immunoblot analysis of the developmental Toxicol 2001; 15: 67–75. expression of individual UDP-glucuronosyltransferase isoenzymes 37 Kakuta Y, Petrotchenko EV, Pedersen LC, Negishi M. The sulfuryl in rat and human-liver microsomes. Mol Pharmacol 1988; 34: transfer mechanism: crystal structure of a vanadate complex of estrogen 729–735. sulfotransferase and mutational analysis. J Biol Chem 1998; 273: 27325– 16 Varin L, Marsolais F, Richard M, Rouleau M. Biochemistry and molecular 27330. biology of plant sulfotransferases. FASEB J 1997; 11: 517–525. 38 Dajani R, Hood AM, Coughtrie MWH. A single amino acid, Glu146, 17 Weinshilboum RM, Otterness DM, Aksoy IA, Wood TC, Her C, governs the substrate specificity of a human dopamine sulfotransferase, Raftogianis RB. Sulfotransferase molecular biology: cDNAs and genes. SULT1A3. Mol Pharmacol 1998; 54: 942–948. FASEB J 1997; 11: 3–14. 39 Brix LA, Barnett AC, Duggleby RG, Leggett B, McManus ME. Analysis of 18 Her C, Kaur GP, Athwal RS, Weinshilboum RM. Human sulfotransferase the substrate specificity of human sulfotransferases SULT1A1 and SULT1C1: cDNA cloning, tissue-specific expression and chromosomal SULT1A3: site-directed mutagenesis and kinetic studies. Biochemistry localization. Genomics 1997; 41: 467–470. 1999; 38: 10474–10479. 19 Sakakibara Y, Yanagishita M, Katafuchi J, Ringer DP, Takami Y, 40 Brix LA, Duggleby RG, Gaedigk A, McManus ME. Structural character- Nakayama T et al. Molecular cloning, expression, and characterization ization of human aryl sulphotransferases. Biochem J 1999; 337: 337– of novel human SULT1C sulfotransferases that catalyze sulfonation 343. of N-hydroxy-2-acetylaminofluorene. J Biol Chem 1998; 273: 33929– 41 Petrotchenko EV, Doerflein ME, Kakuta Y, Pedersen LC, Negishi M. 33935. Substrate gating confers steroid specificity to estrogen sulfotransferase. 20 Parker CR, Falany CN, Stockard CR, Stankovic AK, Grizzle WE. J Biol Chem 1999; 274: 30019–30022. Immunohistochemical localization of dehydroepiandrosterone sulfo- 42 Weinshilboum R. Phenol sulfotransferase inheritance. Cell Mol Neurobiol transferase in human fetal tissues. J Clin Endocrinol Metab 1994; 78: 1988; 8: 27–33. 234–236. 43 Weinshilboum R. Sulfotransferase pharmacogenetics. Pharmac Ther 21 Song WC, Melner MH. Editorial: steroid transformation enzymes as 1990; 45: 93–107. critical regulators of steroid action in vivo. Endocrinology 2000; 141: 44 Weinshilboum R, Aksoy I. Sulfation pharmacogenetics in humans. Chem 1587–1589. Biol Interact 1994; 92: 233–246. 22 Buirchell BJ, Ha¨hnel R. Metabolism of estradiol-17b in human 45 Glatt H, Engelke CEH, Pabel U, Teubner W, Jones AL, Coughtrie MWH endometrium during the menstrual cycle. J Steroid Biochem 1975; 6: et al. Sulfotransferases: genetics and role in toxicology. Toxicol Lett 1489–1494. 2000; 112: 341–348.

The Pharmacogenomics Journal Recent advances in sulfotransferase research MWH Coughtrie 307

46 Price RA, Spielman RS, Lucena AL, van Loon JA, Baidak BL, 66 Steiner M, Bastian M, Schulz WA, Pulte T, Franke KH, Rohring A et al. Weinshilboum RM. Genetic polymorphism for human platelet thermo- Phenol sulphotransferase SULT1A1 polymorphism in prostate cancer: stable phenol sulfotransferase (TS PST) activity. Genetics 1989; 122: lack of association. Arch Toxicol 2000; 74: 222–225. 905–914. 67 Kalow W, Bertilsson L. Interethnic factors affecting drug response. 47 Reiter C, Weinshilboum R. Platelet phenol sulfotransferase activity: Advanc Drug Res 1994; 23: 1–53. correlation with sulfate conjugation of acetaminophen. Clin Pharmacol 68 Nebert DW, Menon AG. Pharmacogenomics, ethnicity, and suscept- Ther 1982; 32: 612–621. ibility genes. Pharmacogenomics J 2001; 1: 19–22. 48 Bonham-Carter SM, Rein G, Glover V, Sandler M, Caldwell J. Human 69 Zhu X, Veronese ME, Iocco P, McManus ME. cDNA cloning and platelet phenol sulphotransferase M and P: substrate specificities and expression of a new form of human aryl sulfotransferase. Int J Biochem correlation with in vivo sulphoconjugation of paracetamol and Cell Biol 1996; 28: 565–571. salicylamide. Br J Clin Pharmacol 1983; 15: 323–330. 70 Engelke CEH, Meinl W, Boeing H, Glatt H. Association between 49 van Loon JA, Weinshilboum RM. Human platelet phenol sulfotransfer- functional genetic polymorphisms of human sulfotransferases 1A1 and ase: familial variation in thermal stability of the TS form. Biochem Genet 1A2. Pharmacogenetics 2000; 10: 163–169. 1984; 22: 997–1014. 71 Dooley TP, Haldeman-Cahill R, Joiner J, Wilborn TW. Expression 50 Jones AL, Roberts RC, Coughtrie MWH. The human phenolsulfotrans- profiling of human sulfotransferase and sulfatase gene superfamilies in ferase polymorphism is determined by the level of expression of the epithelial tissues and cultured cells. Biochem Biophys Res Commun 2000; enzyme protein. Biochem J 1993; 296: 287–290. 277: 236–245. 51 Stanley EL, Hume R, Visser TJ, Coughtrie MWH. Differential expression 72 Carlini EJ, Raftogianis RB, Wood TC, Jin F, Zheng W, Rebbeck TR et al. of sulfotransferase enzymes involved in thyroid hormone metabolism Sulfation pharmacogenetics: SULT1A1 and SULT1A2 allele frequencies during human placental development. J Clin Endocrinol Metab 2001; 86: in Caucasian, Chinese and African–American subjects. Pharmacogenetics 5944–5955. 2001; 11: 57–68. 52 Raftogianis RB, Wood TC, Otterness DM, van Loon JA, Weinshilboum 73 Iida A, Sekine A, Saito S, Kitamura Y, Kitamoto T, Osawa S et al. Catalog RM. Phenol sulfotransferase pharmacogenetics in humans: association of 320 single nucleotide polymorphisms (SNPs) in 20 quinone of common SULT1A1 alleles with TS PST phenotype. Biochem Biophys oxidoreductase and sulfotransferase genes. J Hum Genet 2001; 46: Res Commun 1997; 239: 298–304. 225–240. 53 Raftogianis RB, Wood TC, Weinshilboum RM. Human phenol sulfo- 74 Freimuth RR, Eckloff B, Wieben ED, Weinshilboum RM. Human transferases SULT1A2 and SULT1A1. Genetic polymorphisms, allozyme sulfotransferase SULT1C1 pharmacogenetics: gene resequencing properties, and human liver genotype–phenotype correlations. Biochem and functional genomic studies. Pharmacogenetics 2001; 11: Pharmacol 1999; 58: 605–616. 747–756. 54 Nowell S, Ambrosone CB, Ozawa S, MacLeod SL, Mrackova G, Williams 75 Freimuth RR, Raftogianis RB, Wood TC, Moon E, Kim UJ, Xu J et al. S et al. Relationship of phenol sulfotransferase activity (SULT1A1) Human sulfotransferases SULT1C1 and SULT1C2: cDNA characteriza- genotype to sulfotransferase phenotype in platelet cytosol. Pharmaco- tion, gene cloning, and chromosomal localization. Genomics 2000; 65: genetics 2000; 10: 789–797. 157–165. 55 Ozawa S, Shimizu M, Katoh T, Miyajima A, Ohno Y, Matsumoto Y et al. 76 Yoshinari K, Nagata K, Shimada M, Yamazoe Y. Molecular characteriza- Sulfating-activity and stability of cDNA-expressed allozymes of human tion of ST1C1-related human sulfotransferase. Carcinogenesis 1998; 19: phenol sulfotransferase, ST1A3*1 (213Arg) and ST1A3*2 (213His), both 951–953. of which exist in Japanese as well as Caucasians. J Biochem 1999; 126: 77 Li XY, Clemens DL, Anderson RJ. Sulfation of iodothyronines by human 271–277. sulfotransferase 1C1 (SULT1C1). Biochem Pharmacol 2000; 60: 1713– 56 Li X, Clemens DL, Cole JR, Anderson RJ. Characterization of human liver 1716. thermostable phenol sulfotransferase (SULT1A1) allozymes with 78 Song W-C, Qian YM, Li AP. Estrogen sulfotransferase expression in the 3,30,50-triiodothyronine as the substrate. J Endocrinol 2001; 171: 525– human liver: marked interindividual variation and lack of gender 532. specificity. J Pharmacol Exp Ther 1998; 284: 1197–1202. 57 Raftogianis RB. Human sulfotransferases and cellular response to 79 Rotter JI, Wong FL, Lifrak ET, Parker LN. A genetic component to the estrogens and antiestrogens. Drug Metab Rev 2001; 33:10. variation of dehydroepiandrosterone sulfate. Metabolism 1985; 34: 58 Petrotchenko EV, Pedersen LC , Borchers CH, Tomer KB, Negishi M. The 731–736. dimerization motif of cytosolic sulfotransferases. FEBS Lett 2001; 490: 80 Aksoy IA, Sochorova´ V, Weinshilboum RM. Human liver dehydroepian- 39–43. drosterone sulfotransferase—nature and extent of individual variation. 59 Kester MHA, Kaptein E, Roest TJ, van Dijk CH, Tibboel D, Meinl W et al. Clin Pharmacol Ther 1993; 54: 498–506. Characterization of human iodothyronine sulfotransferases. J Clin 81 Wood TC, Her C, Aksoy I, Otterness DM, Weinshilboum RM. Human Endocrinol Metab 1999; 84: 1357–1364. dehydroepiandrosterone sulfotransferase pharmacogenetics: quantita- 60 Coughtrie MWH, Gilissen RAHJ, Shek B, Strange RC, Fryer AA, Jones PW tive western analysis and gene sequence polymorphisms. J Steroid et al. Phenol sulphotransferase SULT1A1 polymorphism: molecular Biochem Mol Biol 1996; 59: 467–478. diagnosis and allele frequencies in Caucasian and African populations. 82 Nagata K, Yamazoe Y. Pharmacogenetics of sulfotransferase. Annu Rev Biochem J 1999; 337: 45–49. Pharmacol Toxicol 2000; 40: 159–176. 61 Frame LT, Gatlin TL, Kadlubar FF, Lang NP. Metabolic differences and 83 Thomae BA, Eckloff BW, Freimuth RR, Wieben ED, Weinshilboum RM. their impact on human disease—Sulfotransferase and colorectal cancer. Human sulfotransferase SULT2A1 pharmacogenetics: genotype-to- Environ Toxicol Pharmacol 1997; 4: 277–281. phenotype studies. Pharmacogenomics J 2002; 2: 48–56. 62 Bamber DE, Fryer AA, Strange RC, Elder JB, Deakin M, Rajagopal R et al. 84 Ingelman-Sundberg M. Implications of polymorphic cytochrome P450- Phenol sulphotransferase SULT1A1*1 genotype is associated with dependent drug metabolism for drug development. Drug Metab Dispos reduced risk of colorectal cancer. Pharmacogenetics 2001; 11: 2001; 29: 570–573. 679–685. 85 Hodgson J. ADMET—turning chemicals into drugs. Nat Biotechnol 2001; 63 Seth P, Lunetta KL, Bell DW, Gray H, Nasser SM, Rhei E et al. Phenol 19: 722–726. sulfotransferases: hormonal regulation, polymorphism, and age of onset 86 Funk C, Ponelle C, Scheuermann G, Pantze M. Cholestatic potential of of breast cancer. Cancer Res 2000; 60: 6859–6863. troglitazone as a possible factor contributing to troglitazone-induced 64 Zheng W, Xie DW, Cerhan JR, Sellers TA, Wen WQ, Folsom AR. hepatotoxicity: In vivo and in vitro interaction at the canalicular Sulfotransferase 1A1 polymorphism, endogenous estrogen exposure, bile salt export pump (Bsep) in the rat. Mol Pharmacol 2001; 59: well-done meat intake, and breast cancer risk. Cancer Epidem Biomar 627–635. 2001; 10: 89–94. 87 Keppler D, Kamisako T, Leier I, Cui Y, Nies AT, Tsujii H et al. Localization, 65 Chou HC, Lang NP, Kadlubar FF. Metabolic activation of the N-hydroxy substrate specificity, and drug resistance conferred by conjugate derivative of the carcinogen 4-aminobiphenyl by human tissue export pumps of the MRP family. Adv Enzyme Regul 2000; 40: sulfotransferases. Carcinogenesis 1995; 16: 413–417. 339–349.

www.nature.com/tpj Recent advances in sulfotransferase research MWH Coughtrie 308

88 Suzuki H, Sugiyama Y. Transport of drugs across the hepatic sinusoidal 93 Meloche CA, Falany CN. Expression of SULT2B1b in hormonally membrane: sinusoidal drug influx and efflux in the liver. Semin Liver Dis regulated human tissues. Drug Metab Rev 2001; 33: 230. 2000; 20: 251–263. 94 Geese WJ, Raftogianis RB. Biochemical characterization and tissue 89 Kullak-Ublick GA, Stieger B, Hagenbuch B, Meier PJ. Hepatic transport distribution of human SULT2B1. Biochem Biophys Res Commun 2001; of bile salts. Semin Liver Dis 2000; 20: 273–292. 288: 280–289. 90 Qian YM, Sun XJ, Tong MH, Li XP, Richa J, Song WC. Targeted 95 Javitt NB, Lee YC, Shimizu C, Fuda H, Strott CA. Cholesterol and disruption of the mouse estrogen sulfotransferase gene reveals a role of hydroxycholesterol sulfotransferases: identification, distinction from estrogen metabolism in intracrine and paracrine estrogen regulation. dehydroepiandrosterone sulfotransferase, and differential tissue expres- Endocrinology 2001; 142: 5342–5350. sion. Endocrinology 2001; 142: 2978–2984. 91 Song W-C, Qian Y, Sun X, Negishi M. Cellular localization and 96 Ozawa S, Tang Y-M, Yamazoe Y, Kato R, Lang NP, Kadlubar FF. regulation of expression of testicular estrogen sulfotransferase. Endocri- Genetic polymorphism in human liver phenol sulfotransferases nology 1997; 138: 5006–5012. involved in the bioactivation of N-hydroxy derivatives of carcinogenic 92 Borthwick EB, Burchell A, Coughtrie MWH. Differential expression of arylamines and heterocyclic amines. Chem Biol Interact 1998; 109: hepatic estrogen, phenol and dehydroepiandrosterone sulfotransferases 237–248. in genetically-obese diabetic (ob/ob) male and female mice. J Endocrinol 1995; 144: 31–37.

The Pharmacogenomics Journal