This Article Abstract Full Text (PDF) Purchase Article View Shopping Cart Alert me when this article is cited Alert me if a correction is posted Citation Map Services Email this article to a friend Similar articles in this journal Similar articles in PubMed Alert me to new issues of the journal Download to citation manager Request Copyright Permission Citing Articles Citing Articles via HighWire Citing Articles via Google Scholar Google Scholar Articles by Strott, C. A. Search for Related Content PubMed PubMed Citation Articles by Strott, C. A.

Sulfonation and Molecular Action

Section on Steroid Regulation, Endocrinology and Reproduction Research Branch, National Institute of Child Health and Human Development, National Institutes of Health, Bethesda, Maryland 20892-4510

Correspondence: Address all correspondence and requests for reprints to: Charles A. Strott, M.D., Building 49, Room 6A36, National Institutes of Health, Bethesda, Maryland 20892-4510. E-mail: chastro{at}mail.nih.gov

	Abstract

Top Abstract I. Introduction II. The Universal Sulfonate... III. Sulfotransferases IV. Sulfonation of... V. Sulfonation and Thyroid... VI. Sulfonation of... VII. Sulfonation of Peptide... VIII. Sulfonation of... IX. Summary and Future... References

 
The sulfonation of endogenous molecules is a pervasive biological phenomenon that is not always easily understood, and although it is increasingly recognized as a function of fundamental importance, there remain areas in which significant cognizance is still lacking or at most minimal. This is particularly true in the field of endocrinology, in which the sulfoconjugation of hormones is a widespread occurrence that is only partially, if at all, appreciated. In the realm of steroid/sterol sulfoconjugation, the discovery of a novel gene that utilizes an alternative exon 1 to encode for two sulfotransferase isoforms, one of which sulfonates cholesterol and the other pregnenolone, has been an important advance. This is significant because cholesterol sulfate plays a crucial role in physiological systems such as keratinocyte differentiation and development of the skin barrier, and pregnenolone sulfate is now acknowledged as an important neurosteroid. The sulfonation of thyroglobulin and thyroid hormones has been extensively investigated and, although this transformation is better understood, there remain areas of incomplete comprehension. The sulfonation of catecholamines is a prevalent modification that has been extensively studied but, unfortunately, remains poorly understood. The sulfonation of pituitary glycoprotein hormones, especially LH and TSH, does not affect binding to their cognate receptors; however, sulfonation does play an important role in their plasma clearance, which indirectly has a significant effect on biological activity. On the other hand, the sulfonation of distinct neuroendocrine peptides does have a profound influence on receptor binding and, thus, a direct effect on biological activity. The sulfonation of specific extracellular structures plays an essential role in the binding and signaling of a large family of extracellular growth factors. In summary, sulfonation is a ubiquitous posttranslational modification of hormones and extracellular components that can lead to dramatic structural changes in affected molecules, the biological significance of which is now beginning to be appreciated.

I. Introduction

II. The Universal Sulfonate Donor Molecule

A. General

B. PAPS synthesis

C. Cloning and characterization of PAPS synthases

D. Regulation of PAPS availability

E. Tissue levels of PAPS

III. Sulfotransferases

A. General

B. Classification

C. Biochemistry

D. Structure

IV. Sulfonation of Steroids/Sterols

A. General

B. Novel steroid/sterol sulfotransferase subfamily

C. Transcriptional regulation of steroid sulfotransferases

D. Structural analysis of steroid sulfotransferases

E. Biology

V. Sulfonation and Thyroid Function

A. General

B. Sulfonation of thyroglobulin

C. Sulfonation of thyroid hormones

D. Sulfonation of TSH

E. Sulfotransferases

VI. Sulfonation of Catecholamines

A. General

B. Sources of catecholamine sulfates

C. Catecholamine sulfotransferases

D. Physiology and clinical significance of catecholamine sulfonation

VII. Sulfonation of Peptide Hormones

A. General

B. Neuroendocrine peptides

C. Peptidyl sulfotransferases

D. Glycoprotein hormones

E. N-Acetylgalactosamine-4-O-sulfotransferase

VIII. Sulfonation of Extracellular Structures and Signaling

A. General

B. Heparan sulfate

C. Sulfotransferases

IX. Summary and Future Directions

	I. Introduction

Top Abstract I. Introduction II. The Universal Sulfonate... III. Sulfotransferases IV. Sulfonation of... V. Sulfonation and Thyroid... VI. Sulfonation of... VII. Sulfonation of Peptide... VIII. Sulfonation of... IX. Summary and Future... References

 
THE BIOTRANSFORMATION OF molecules by sulfonation is a basic metabolic route of primary importance. Sulfonation is the transfer of a sulfonate group (SO₃^-1) from the universal sulfonate donor 3'-phosphoadenosine 5'-phosphosulfate (PAPS) to an appropriate acceptor molecule. Sulfonated compounds comprise a remarkable array of substances, ranging in molecular weight from less than 10³ to greater than 10⁶, that undergo striking changes in their physicochemical properties upon the addition of the highly charged sulfonate group (1). Sulfonation increases water solubility and can lead to conformational changes in both low- and high-molecular-weight molecules; lipophilic molecules are converted to amphiphiles and, with a pK_a near 1.5, sulfonates remain fully ionized at any pH found in biological systems (1).

Interestingly, two of the more prominent conjugating systems involving the nonmetal elements phosphorous and sulfur sit side by side in the periodic table. In metazoan physiology, phosphorylation and sulfonation are ubiquitous phenomena carried out in all organ systems. Because of the broad role played by phosphorylation in regulatory mechanisms, particularly involving enzymes, signal transduction pathways, and transcription, it continues to receive extensive coverage. The importance of sulfonation, however, is less well appreciated, despite the fact that it is absolutely essential for normal growth and development as well as maintenance of the internal milieu. Sulfonated macromolecules such as glycosaminoglycans and proteoglycans are involved in cell surface and connective tissue structures. The highly acidic and hydrophilic glycosaminoglycans have a major influence on tissue hydration, elasticity, and cation composition (2). Furthermore, they participate directly in high-affinity binding to extracellular matrix proteins, growth factors, enzymes, and cell surface receptors (3, 4, 5), and they engage in transmembrane signaling (6). Sulfonation of tyrosine residues is a prevalent posttranslational modification of many secretory and membrane proteins and peptides that may significantly influence functionality (7, 8). Sulfate moieties in sugar residues of glycoprotein hormones have a significant influence on biological activity (9, 10). Sulfoglycolipids such as sphingolipids and galactoglycerolipids are abundant in myelin as well as spermatozoa, kidney, and small intestine (11) and have been implicated in a variety of physiological functions through their interactions with extracellular matrix proteins, cellular adhesive receptors, blood coagulation systems, complement activation systems, and cation transport systems (12). Sulfonation also has a significant role in the biotransformation of many endogenous low-molecular-weight compounds, including catecholamines (13), iodothyronines (14), and vitamin C (15). Likewise, sulfonation is an important modification of cholesterol (16) and its derivatives, bile acids (17), vitamin D (18, 19), and steroids (20).

The principal aim of this review is to elaborate on the mechanism and function of sulfonation in a number of the basic systems noted above, especially those related to endocrinology. The primary intent is to summarize the current status of sulfoconjugation in these systems, particularly as this modification applies to molecular mechanisms operative in human endocrine physiology. One area that will be touched on in the review involves the extracellular matrix and proteoglycans, the latter being a term not exactly in the working lexicon of most endocrinologists. However, sulfonated proteoglycans and their role in cell signaling is a subject of great importance, especially regarding the molecular action of specific growth factors. In the final analysis, it is hoped that the reader will come away from the review with an enhanced appreciation for the pervasiveness as well as importance of sulfonation in mammalian biology.

	II. The Universal Sulfonate Donor Molecule

Top Abstract I. Introduction II. The Universal Sulfonate... III. Sulfotransferases IV. Sulfonation of... V. Sulfonation and Thyroid... VI. Sulfonation of... VII. Sulfonation of Peptide... VIII. Sulfonation of... IX. Summary and Future... References

 
A. General
The elements carbon, nitrogen, and sulfur are made available to organisms chiefly in the form of inorganic compounds. For sulfur, the inorganic compound is sulfate. Furthermore, as in the case of carbon dioxide formation and nitrogen fixation, the utilization of sulfate also requires metabolic activation to a form that can be reduced to sulfide, which is then used in the production of cysteine and methionine needed for the synthesis of protein (21). Multicellular organisms, however, are unable to reduce sulfate to sulfide, just as they cannot reduce carbon dioxide to carbohydrate and nitrate to ammonia. Thus, the animal kingdom must rely on plants and bacteria to provide the reduced forms of these elements (21). In the case of sulfur, the activated sulfate compound is PAPS (22, 23). The importance of PAPS is that it serves not only as a substrate for reduction by bacteria and plants but also as the active agent for sulfonate esterification, a process carried out by all organisms. PAPS thus serves as the universal sulfonate donor molecule required for all sulfonation reactions (24); in mammals, all tissues are able to carry out the synthesis of PAPS (25).

B. PAPS synthesis
PAPS synthesis requires a ready supply of sulfate, which is available from the diet as well as catabolism of proteins and sugar sulfates (1). Abnormalities resulting from sulfate deficiency are theoretically possible but are distinctly unusual under normal circumstances, because plasma levels of sulfate are resistant to manipulation (1). There are, however, genetic disorders involving the cellular uptake of sulfate by a carrier-mediated transport (26). Achondrogenesis type 1B is a recessively inherited chondrodysplasia characterized by extremely poor skeletal development and perinatal death (27). Atelosteogenesis type II is a recessively inherited neonatally lethal chondrodysplasia characterized by defective uptake of inorganic sulfate and insufficient sulfonation of macromolecules (28). These two genetic disorders, and a third recessively inherited but nonlethal chondrodysplasia (diastrophic dysplasia), are caused by mutations in the diastrophic dysplasia sulfate transporter gene located on the long arm of chromosome 5 (29). In the three conditions, impaired sulfate transport across cell membranes results in undersulfonation of cartilaginous proteoglycans. Phenotypic severity correlates with the underlying sulfate transporter gene mutations: homozygosity or compound heterozygosity for stop codons, or transmembrane domain substitutions, usually result in achondrogenesis type 1B, whereas other structural or regulatory mutations result in less severe phenotypes (29). Interestingly, Pendred’s syndrome (sporadic goiter with impaired iodine organification and congenital sensorineural deafness) is an autosomal recessive disorder caused by a mutated gene located on chromosome 7q22–31.1. The Pendred syndrome gene, which produces a transcript that is expressed in the thyroid, the inner ear, and the kidney (30, 31), was originally thought to be a sulfate transporter gene (32), because the predicted protein, pendrin, has high homology to several sulfate transporters found in yeast, plants, and animals (30). Subsequently, however, it was shown that pendrin, which belongs to a large family of anion transporters (32), functions as a transporter of chloride and iodide but not sulfate (33). Nevertheless, it is of interest that one of pendrin’s closest mammalian relatives is the protein encoded by the diastrophic dysplasia sulfate transporter gene (34).

The activation of inorganic sulfate to form PAPS results from the concerted action of two enzyme systems (35, 36, 37), which in animal species is carried out by a bifunctional protein (38). The first step (Fig. 1) is catalyzed by ATP sulfurylase and involves the reaction of inorganic sulfate with ATP to form adenosine 5'-phosphosulfate (APS) and inorganic phosphate. This reaction results in the formation of a high-energy phosphoric-sulfuric acid anhydride bond that is the chemical basis for sulfate activation (39). The second step (Fig. 1) is catalyzed by APS kinase and involves the reaction of APS with another molecule of ATP to form PAPS and ADP. Unlike ATP sulfurylase, APS kinase is not involved in the activation of sulfate, and its raison d’être is not well understood (39). In an interesting departure from sulfonation reactions, it was found that PAPS could also serve as a phosphate donor in protein phosphorylation. That is, the 3'-phosphate group of PAPS was transferred to a serine residue in an 85-kDa membrane protein (40). Furthermore, this phenomenon was carried out by 11 different tissues examined, suggesting the existence of a novel widespread form of phosphorylation (40). This intriguing in vitro finding notwithstanding, its physiological significance remains to be elucidated.

View larger version (17K): [in this window] [in a new window]   Figure 1. Catalytic reactions in the formation of PAPS.

C. Cloning and characterization of PAPS synthases
PAPS synthase exists as two isozymes encoded by genes located on separate chromosomes. PAPS synthase 1 has been cloned from several species, including the marine worm (43), mouse (44), human (45, 46, 47), Drosophila (48), and guinea pig (49). The gene for human PAPS synthase 1 is located on chromosome 4q25–26 (50). Human PAPS synthase 1 is 98% and 95% identical with mouse and guinea pig PAPS synthase 1, respectively, indicating that it is a highly conserved protein. The catalytic domain of APS kinase is located in the amino-terminal region, whereas the ATP sulfurylase domain is in the carboxy-terminal section of this bifunctional protein (45). The division between the two domains, as determined for human PAPS synthase 1, is at the junction of amino acids 226 and 227 of the 624-amino-acid protein (cf. Fig. 2; author’s unpublished data). Interestingly, there are two completely conserved nucleotide-binding motifs, one in each domain (cf. Fig. 2).

View larger version (77K): [in this window] [in a new window]   Figure 2. Amino acid alignment of human PAPS synthase 1 (hPAPSS1), human PAPS synthase 2a (hPAPSS2a), and human PAPS synthase 2b (hPAPSS2b). The dashed box defines the APS kinase domain, and the external solid box outlines the ATP sulfurylase domain. The P-loop motif in the APS kinase domain and the HxGH motif in the ATP sulfurylase domain are boxed and labeled. The GMALP sequence in the ATP sulfurylase domain of hPAPSS2b is shown by reverse type. Identities are shaded.

After the cloning of PAPS synthase 1, a second isozyme (PAPS synthase 2) was cloned for human (56), mouse (56, 57), and guinea pig species (49). The gene for human PAPS synthase 2 has been localized to chromosome 10q23–24 (56). The mouse and human PAPS synthase 2 isozymes were discovered through investigation of specific developmental dwarfing disorders, i.e., brachymorphism in mice and a form of spondyloepimetaphyseal dysplasia in humans. In both human and mouse species, the PAPS synthase 1 and 2 proteins are 77% identical. An interesting feature of the human and mouse PAPS synthase 2 genes is the occurrence of alternative splicing (58), which results in the formation of two variants (2a and 2b) distinguished by the presence or absence of a five-amino-acid segment (GMALP) in the ATP sulfurylase domain of the protein (cf. Fig. 2). The catalytic activity of the human PAPS synthase 2a splice variant is modestly but significantly less (30%) than that of the 2b variant (49).

The physiological significance of the two genes encoding for related proteins that carry out an identical function, i.e., synthesis of the essential sulfonate donor molecule, PAPS, is presently unclear. Interestingly, the catalytic activity of the PAPS synthase 2 variants is 10- to 15-fold higher than that for PAPS synthase 1 (49). It is difficult to imagine that this represents a backup system, because the human PAPS synthase 2 variants appear to be expressed in a tissue-specific manner in contrast to the ubiquitously expressed PAPS synthase 1 (49). What then is the relationship of PAPS synthase 1 to PAPS synthase 2? It was recently reported that PAPS synthase 1 has a nuclear localization in mammalian cells; on the other hand, PAPS synthase 2 has a nuclear localization only when coexpressed with PAPS synthase 1 (59). It was suggested that PAPS synthase 1 localizes to the nucleus in most cells, whereas PAPS synthase 2 localizes to the cytoplasm in tissues in which PAPS synthase 1 levels are low. Interestingly, it was determined that nuclear targeting of PAPS synthase 1 requires the 21-amino-acid segment from the catalytically dispensable amino terminus (59). This intriguing hypothesis notwithstanding, it is difficult to understand the physiological significance. Although some sulfonation may necessarily occur in the nucleus, e.g., estrogen sulfotransferase in the guinea pig and rat has a nuclear localization (60, 61), it would appear that the cytoplasmic compartment is the major site of sulfonation involving secretory proteins, proteoglycans, glycosaminoglycans, galactoglycerolipids, sphingolipids, and a myriad of small-molecular-weight compounds such as hormones and neurotransmitters, as well as drugs and xenobiotics.

Human PAPS synthase 2 was discovered during a search for the genetic basis of a developmental abnormality resulting in a form of spondyloepimetaphyseal dysplasia that presents with a skeletal phenotype involving the spine and long bones. This recessive dwarfing disorder is caused by a nonsense mutation located in the ATP sulfurylase domain of PAPS synthase 2 (56). Brachymorphism, a comparable dwarfing abnormality occurring in mice, is due to a mutation in the syntenic gene for PAPS synthase 2; in this case, there is a missense mutation in the APS kinase domain of the protein (57). It was baffling that in the genetic disorder involving human PAPS synthase 2, which produces the osteochondrodysplasia phenotype, the cartilage-specific defect occurs despite the coexpression of PAPS synthase 1 in cartilaginous tissue (56). In fact, in adult human cartilage, PAPS synthase 1 appears to be the dominant isoform (49). This apparent enigma would seem to be resolved by examination of cartilage from guinea pigs as an animal model. Similar to humans, cartilage from mature animals predominantly expresses PAPS synthase 1. In contrast, however, PAPS synthase 1 expression is relatively low in the cartilage of immature guinea pigs including the growth plate of long bones, whereas PAPS synthase 2 is the vigorously expressed isozyme (49).

Defects involving the PAPS synthase 1 gene have not been reported, presumably because it is ubiquitously expressed, in contrast to PAPS synthase 2; furthermore, it is the predominant, if not sole, isoform expressed in the central nervous system and bone marrow, in which genic mutations are likely to be embryologically lethal.

D. Regulation of PAPS availability
The fact that PAPS is such a strategic biological molecule makes understanding the molecular mechanisms involved in regulating its availability of vital importance. Information in this area, however, is presently rather limited. Little is known, for instance, regarding the transcriptional regulation of the PAPS synthase genes, or the stability of mRNA encoding PAPS synthase proteins, or specific tissue turnover of PAPS synthase proteins. Proximal promoter regions of the genes for human PAPS synthase 1 and 2, which contain neither a TATAAA nor a CCAAT box, have been identified and, in each case, found to be under the influence of the Sp1 family of transcription factors (62, 63). Although these findings are of some interest, they nevertheless represent only incipient observations on the transcriptional regulation of these important genes.

The principal PAPS-degrading enzymes in animal tissues are nucleotidases and sulfohydrolases (23, 64, 65). A 3'- nucleotidase converts PAPS to APS and PAP (desulfonated PAPS) to AMP by hydrolysis of the 3'-phosphate bond (66, 67). A 5'-nucleotidase hydrolyzes the 5'-phosphosulfate bond of PAPS and APS (68, 69, 70). A sulfohydrolase converts PAPS to PAP and APS to AMP (71, 72, 73). PAPS-metabolizing enzymes appear to be widely distributed and are localized to both particulate and soluble tissue fractions (23, 64, 65). In animal species, none of these enzymes has been suitably purified and adequately characterized, and except for the cloning of a 3',5'-bisphosphate nucleotidase in yeast, no PAPS-metabolizing enzyme has been cloned (74, 75). Although it can be postulated that such enzymes are involved in the regulation of tissue concentrations of PAPS, there are no pertinent experimental results that address this issue. It has been suggested that high tissue concentrations of APS might inhibit the APS kinase reaction (76). This suggestion is based on a study of the fungus Penicillium chrysogenum (77), a species in which ATP sulfurylase and APS kinase are on separate polypeptide chains and channeling of APS from ATP sulfurylase to APS kinase does not occur (78). On the other hand, bifunctional PAPS synthase does demonstrate channeling of APS (42); therefore, it would seem unlikely that exogenous APS would significantly inhibit the APS kinase reaction, a conclusion supported by recent experimental evidence (49).

E. Tissue levels of PAPS
There is a limited amount of information available regarding organ concentrations of PAPS in different species under presumably basal metabolic conditions. For instance, in rat liver and kidney, PAPS levels range from 60–160 and 40–50 nmol/g tissue, respectively (79, 80, 81, 82, 83). The concentration of PAPS in lung, intestine, and brain, as well as in liver and kidney tissues of the rat, mouse, hamster, rabbit, and dog, indicates that for each species the highest level is in the liver and ranges from approximately 15 nmol/g tissue for the dog to approximately 80 nmol/g tissue for the rat (84); the concentration of PAPS in various other tissues varies between 6 and 20 nmol/g tissue, with no significant species or sex differences (84). The level of PAPS in several human tissues, including liver, lung, kidney, ileum, and colon, ranges from approximately 4 nmol/g tissue in the lung to approximately 23 nmol/g tissue in the liver (85). The level of PAPS in human fetal liver is approximately 10 nmol/g tissue vs. approximately 23 nmol/g tissue in adult liver and approximately 4 nmol/g tissue in the placenta (86).

Tissue (mostly liver) levels of PAPS have been monitored during various manipulations, e.g., during a sulfur-deficient diet, treatment with inhibitors of energy metabolism, and the administration of xenobiotics to accelerate sulfonation. Such studies, performed primarily in rats, demonstrate the ability to experimentally modulate hepatic levels of PAPS. There are, however, no published reports on correlating production and metabolism of PAPS with a sulfonation step for any tissue. Thus, the influence of PAPS availability on a specific sulfoconjugation reaction under strict experimental conditions, involving either endocrine or metabolic manipulations, remains to be investigated.

	III. Sulfotransferases

Top Abstract I. Introduction II. The Universal Sulfonate... III. Sulfotransferases IV. Sulfonation of... V. Sulfonation and Thyroid... VI. Sulfonation of... VII. Sulfonation of Peptide... VIII. Sulfonation of... IX. Summary and Future... References

 
A. General
Sulfonation reactions are usually classified by the acceptor group involved in sulfoconjugation, e.g., O-sulfonation (ester), N-sulfonation (amide), and S-sulfonation (thioester) (1). O-Sulfonation involves an alcohol group and can occur with diverse, relatively small endogenous compounds such as catecholamines, steroids, thyroid hormones, and vitamins. Similarly, macromolecules such as glycosaminoglycans, proteoglycans, proteins, and galactoglycerolipids are subject to O-sulfonation. In general, O-sulfonation represents the dominant cellular sulfonation reaction. N-Sulfonation, although relatively less prominent than O-sulfonation (25), is nevertheless a crucial reaction in the modification of carbohydrate chains in macromolecules such as heparin and heparan sulfate proteoglycans (87, 88). N-Sulfonation is also involved in the metabolism of xenobiotics (89). S-Sulfonation is not relevant to this review.

B. Classification
Sulfotransferases can be divided into two classes based upon whether they are soluble or membrane-associated proteins. Soluble or cytosolic sulfotransferases sulfonate a wide variety of endogenous compounds including hormones and neurotransmitters as well as drugs and xenobiotics. Membrane-associated sulfotransferases are located in the trans-Golgi complex, where they are involved in the posttranslational modification of macromolecules such as secretory proteins and glycosaminoglycans. There is an effort afoot to establish a standard nomenclature for cytosolic sulfotransferases stemming from an international workshop on sulfation held a few years ago in Drymen, Scotland. The term SULT was adopted as an abbreviation for cytosolic sulfotransferases and their genes (90). Although this is not as yet the "official" nomenclature, it will nevertheless be the general form used in this review.

1. Cytosolic sulfotransferases.
This class of sulfotransferases embodies an ever-enlarging superfamily of enzymes that catalyze the sulfonation of relatively small endobiotics and exobiotics. In mammals, at least 44 cytosolic sulfotransferases have been identified that comprise five SULT families that share less than 40% similarity with each other (91). Of these five SULT families, the first two represent the largest and most widely examined families. The SULT1 family consists of sulfotransferases that transfer sulfonate to phenolic drugs and catecholamines (SULT1A; human chromosome 16), estrogenic steroids (SULT1E) and thyroid hormones (SULT1B; human chromosome 4), and xenobiotics (SULT1C; human chromosome 2). The SULT2 family members primarily sulfonate neutral steroids (SULT2A) and sterols (SULT2B; human chromosome 19). Enzymes within the SULT3 family catalyze the formation of sulfamates, whereas the SULT4 and SULT5 families (human chromosome 22) consist of unique cDNAs found in the DNA database and have not yet been adequately characterized (91). Chromosomal clustering of subfamily members suggests that the emergence of isozymes has probably resulted from gene duplication.

Recently, brain-specific sulfotransferases that have less than 36% amino acid sequence identities to other known cytosolic sulfotransferases have been cloned from human (92, 93), rat (92), and mouse (93). These novel, highly conserved (98% identical) cytosolic sulfotransferases, whose endogenous substrate(s) are unknown, appear to belong to a family of cytosolic sulfotransferases distinct from the two major families noted above, and they have been designated SULT4A1 (94).

The biotransformation of endogenous substrates and xenobiotics by sulfonation is a major metabolic reaction that has two possible consequences, i.e., activation or inactivation of a biological effect. An example of the former involving an endobiotic is the conversion of pregnenolone to pregnenolone sulfate, which is a potent neuroexcitatory agent by virtue of its antagonistic action on the -aminobutyric acid (GABA)_A receptor that regulates chloride channels (95). Another example in which sulfoconjugation leads to activation, in this case involving a xenobiotic, is the drug minoxidil, a potent vasodilator and trichogen whose active form is minoxidil sulfate (96). On the other hand, inactivation of a biological effect can also be produced by sulfoconjugation. For instance, the genomic action of steroid hormones is inhibited by sulfoconjugation because the sulfates of steroid hormones are unable to bind to their cognate nuclear receptors (97). With a few notable exceptions (such as minoxidil), sulfoconjugation is an important mechanism in the inactivation and excretion of drugs and xenobiotics (1, 25). Nevertheless, whereas xenobiotics are usually detoxified by sulfonation, it is noteworthy that a number of compounds (procarcinogens) are converted into highly reactive intermediates by sulfonation and can then act as chemical carcinogens and mutagens by covalently binding to DNA (98, 99, 100). The O-sulfonated product of tamoxifen, the drug widely used in the treatment of breast cancer, is a hepatocarcinogen in rats but not in humans (101), although there is a suggestion that tamoxifen can be genotoxic in humans (102).

2. Golgi-associated sulfotransferases.
This class of sulfotransferases is primarily concerned with the posttranslational modification of carbohydrates, peptides, and proteins. Carbohydrate sulfotransferases are resident transmembrane enzymes of the Golgi network that recognize sugar residues attached to lipids and proteins passing through the secretory pathway (5). In mammals, 32 different carbohydrate sulfotransferases have been cloned and characterized to date (103). Importantly, carbohydrate sulfotransferases are stereoselective and exhibit strict substrate specificities.

Carbohydrates attached to lipids and proteins display complex and heterogeneous structures, and the addition of a sulfonate group can transform a common structural motif of a carbohydrate into a unique recognition site for a specific receptor or lectin (9). One consequence relates to the control of the circulatory half-life of glycoprotein hormones (10). Extracellular sugars are also frequently sulfonated and as a result can mediate numerous highly specific molecular- recognition events such as the binding of growth factors (87).

For many proteins, tyrosine sulfonation is important for biological activity and correct cellular processing (104). Sulfonation is the most abundant posttranslational modification of tyrosine residues involving many soluble and membrane proteins transiting the secretory path (7). An example of tyrosine sulfonation being essential for biological activity is that of cholecystokinin (CCK), which is 250 times more potent in the sulfonated form than in the unconjugated form (105). Tyrosylprotein sulfotransferases are membrane-bound residents of the trans-Golgi network (8), and, to date, two isozymes have been identified and characterized (106, 107).

C. Biochemistry
Sulfonation refers to the transfer of an SO₃^-1 group, whereas sulfation refers to the transfer of an SO₄^-2 group; thus, the products of sulfonation should correctly be referred to as sulfonates (1). However, because transfer of an SO₃^-1 group to a hydroxyl acceptor creates an SO₄ moiety (Fig. 3), sulfonated products have been incorrectly termed "sulfates", a nomenclature that is deeply entrenched. In the general scheme of sulfonation, the sulfonate acceptor (ROH) and donor (PAPS) molecules bind to a sulfotransferase with subsequent release of the sulfonated product and desulfonated PAPS, i.e., PAP (Fig. 3). Sulfotransferases are characteristically high-affinity and low-capacity enzymes. As a consequence, the activity of sulfotransferases is 2–3 orders of magnitude slower than that of phosphotransferases (1). Hydrolysis of the sulfonated product and regeneration of the substrate as a free alcohol complete the sulfurylation cycle, a process carried out by members of the sulfohydrolase or sulfatase gene family (108). Although sulfohydrolysis is extremely important in the overall sulfurylation scheme, this aspect of the sulfurylation cycle will not be covered in this review.

View larger version (12K): [in this window] [in a new window]   Figure 3. Sulfonate ester formation and hydrolysis.

	IV. Sulfonation of Steroids/Sterols

Top Abstract I. Introduction II. The Universal Sulfonate... III. Sulfotransferases IV. Sulfonation of... V. Sulfonation and Thyroid... VI. Sulfonation of... VII. Sulfonation of Peptide... VIII. Sulfonation of... IX. Summary and Future... References

 
A. General
The sulfonation of steroids has received considerable attention during the past decade, largely as a result of cDNA cloning of sulfotransferases from a variety of mammalian species. Because this subject was summarized as recently as 5 yr ago (116, 117), this review will primarily focus on significant developments that have occurred during the interim period. An important advance has been the cloning of a unique hydroxysteroid sulfotransferase subfamily with distinctly novel expression and substrate specificity. A substantial payoff of recombinant DNA technology has been the ability to produce large quantities of a protein for characterization studies and crystallization, as well as the ability to produce altered versions of a specific protein for structure/function analyses. The latter has been an active area of investigation in sulfotransferase research with significant progress in establishing specific structural principles. Additionally, there have been further developments regarding a multiplicity of biological roles attributable to steroid sulfates, as well as the enzymes responsible for their generation.

B. Novel steroid/sterol sulfotransferase subfamily
The cloning of a unique hydroxysteroid sulfotransferase subfamily in human (118) and mouse (119) species has significantly moved the steroid sulfotransferase field forward. The novel human hydroxysteroid sulfotransferase gene (SULT2B1) was discovered during a search of an expressed sequence tag database using a probe containing a highly conserved sulfotransferase sequence (118). The search yielded an expressed tag located at the 3'-end of a clone isolated from a human placental cDNA library by the I.M.A.G.E. Consortium that led to the cloning of two cDNAs designated SULT2B1a and SULT2B1b. These subtypes, which result from the use of an alternative exon 1, thus encode for proteins that differ only at their amino termini. The SULT2B1 gene maps to chromosome 19q13.4, approximately 500 kb telomeric to the location of the related SULT2A1 gene (118).

SULT2A1, the prototypical human hydroxysteroid sulfotransferase (120, 121), is commonly referred to as dehydroepiandrosterone (DHEA) sulfotransferase, because DHEA is considered the major substrate. SULT2A1, however, has broad substrate specificity and will sulfonate a wide variety of steroids and sterols, in addition to DHEA, involving hydroxyl groups at different carbon locations and with different spatial orientations. A 3-hydroxyl group (androsterone and bile acids), a 3ß-hydroxyl group (DHEA and pregnenolone), a 17ß-hydroxyl group (testosterone and estradiol), and a phenolic hydroxyl group (estradiol and estrone) are sulfonated by human SULT2A1 (122, 123, 124, 125, 126). Although human SULT2A1 preferentially sulfonates DHEA, it negligibly sulfonates cholesterol, which is the reverse of the case with human SULT2B1 (127). Similar to SULT2A1, the SULT2B1 isoforms will sulfonate pregnenolone; however, in contrast to SULT2A1, they do not sulfonate androsterone, bile acids, testosterone, estrogens, or cortisol (118, 127, 128, 129). In more recent detailed studies, the human SULT2B1 isoforms demonstrate striking differences in substrate specificity. That is, SULT2B1a shows a distinct preference for pregnenolone as a substrate, whereas cholesterol is only minimally sulfonated (Fig. 4). On the other hand, SULT2B1b vigorously sulfonates cholesterol, whereas pregnenolone is less avidly metabolized (Fig. 4). Additionally, in contrast to SULT2A1, which also does not sulfonate cholesterol, neither SULT2B1 subtype sulfonates DHEA efficiently (Fig. 4). These results suggest that the human SULT2A1 and SULT2B1 isozymes have selective physiological roles and that the SULT2B1b isoform functions as a true cholesterol sulfotransferase.

View larger version (16K): [in this window] [in a new window]   Figure 4. Saturation analyses of bacterially overexpressed and affinity-purified human hydroxysteroid sulfotransferase SULT2B1a (A), SULT2B1b (B), and SULT2A1 (C). Substrates (µM), PAPS cofactor (0.1 mM), and purified proteins (0.1–4.0 µg) were incubated in Tris buffer (pH 7.5) containing MgCl2 (5 mM), hydroxypropyl-ß-cyclodextrin (0.2 mM), and ethanol (4%) at 37 C for 5 min, and the sulfonated products were isolated by thin-layer chromatography. Each point represents the average of duplicate determinations.

View larger version (23K): [in this window] [in a new window]   Figure 5. Sulfotransferase activity of bacterially overexpressed and affinity-purified human hydroxysteroid sulfotransferase, SULT2B1b, using cholesterol or modified-cholesterol as substrate (20 µM). Results are expressed as the amount of modified-cholesterol sulfate formed (nmol/min·mg protein) relative to the amount of cholesterol sulfate formed (nmol/min·mg protein). [Reproduced by permission of N. B. Javitt, unpublished data.]

View larger version (46K): [in this window] [in a new window]   Figure 6. Amino acid sequence alignment of human (h) SULT2A1, SULT2B1a, and SULT2B1b. The unique amino- and carboxy-terminal ends of hSULT2B1a and hSULT2B1b are outlined by dashed boxes. Solid boxes locate conserved residues important in cofactor binding. Arrowheads locate the highly conserved nucleotide-binding motif (PSB) as well as the 5' phosphate (5'PB)- and 3' phosphate (3'PB)-interacting regions. Identities are shaded.

As determined by RT-PCR, the human SULT2A1 and SULT2B1 genes are differentially expressed (127, 132). Human SULT2A1 is robustly expressed in steroidogenic organs (adrenal and ovary), androgen-dependent tissue (prostate), tissues of the alimentary tract (stomach, small intestine, and colon), and the liver; however, it is, notably, not expressed in skin (127). The RT-PCR expression patterns of SULT2B1a and SULT2B1b indicate that SULT2B1b is more widely expressed than SULT2B1a. Importantly, the SULT2B1 subtypes, in contrast to SULT2A1, are vigorously expressed in skin (127). The importance of the observation on the selective expression of SULT2B1 in skin lies in the fact that sulfonation of cholesterol is an essential metabolic step during normal skin development and creation of the barrier (133, 134, 135). Differentiation of normal human epidermal keratinocytes is accompanied by an accumulation of cholesterol sulfate, which is accounted for by an increase in cholesterol sulfotransferase activity (136).

Cloning of a mouse SULT2B1b, which is 71% identical with human SULT2B1b, has also been reported (119). Furthermore, similar to human SULT2B1b, the mouse ortholog has a preference for cholesterol (author’s unpublished data). Recently, the isolation of a sulfotransferase from rat skin was reported to be active against substrates with a ₅ double bond such as cholesterol, whereas androgens, estrogens, corticosteroids, simple phenols, and 3,4-dihydroxyphenylalanine did not serve as substrates (137). The rat cholesterol SULT has an apparent molecular weight of 40,000, which compares with a calculated molecular weight of 38,404 for mouse SULT2B1b and 41,304 for human SULT2B1b. For comparison, human, mouse, and rat SULT2A1 proteins have molecular weights of 33,777; 33,342; and 32,961, respectively. Thus, based on molecular weight and substrate specificity, the recently isolated rat sulfotransferase would appear to belong to the SULT2B1 subfamily (131A ).

C. Transcriptional regulation of steroid sulfotransferases
Although a large number of steroid sulfotransferase genes have now been cloned in several species, there has been little information forthcoming regarding their transcriptional regulation, with the exception of rat DHEA sulfotransferase (SULT2A1), a subject on which two papers have emerged (138, 139). Rat SULT2A1 is selectively manifest in liver, in which expression is strongly repressed by androgens (140). In this regard, hepatocyte nuclear factor-1 (HNF1) and CCAAT/enhancer-binding protein (C/EBP) response elements play pivotal roles (138). Regarding androgen repression, a negative androgen response region in the rat SULT2A1 promoter has been mapped. Androgenic repression of the rat SULT2A1 gene requires the presence of OCT-1 and C/EBP elements that map to specific promoter locations. Furthermore, the negative androgenic regulatory effect may be exerted indirectly through transcriptional interference of OCT-1 and C/EBP rather than via a direct interaction of the androgen receptor with DNA (138).

Human SULT2A1 sulfonates bile acids (122). Thus, it is of interest that the bile acid chenodeoxycholic acid is a strong inducer of the rat SULT2A1 gene (139). Furthermore, the inducing effect is mediated through the bile acid-activated farnesoid X receptor (FXR), a member of the nuclear receptor superfamily. Ligand-activated FXR forms a heterodimer with the 9-cis retinoic acid receptor (RXR) and binds to a defined FXR/RXR promoter position to regulate rat SULT2A1 gene expression (139). The same tissues in the rat (liver, small intestine, colon, and adrenal cortex) express SULT2A1 and FXR, and as a result of bile acid sulfonation, SULT2A1 may play a significant role in cholesterol removal (139).

D. Structural analysis of steroid sulfotransferases
1. Crystallography.
The first steroid sulfotransferase structure to be solved was that of mouse estrogen sulfotransferase (SULT1E1; Ref. 109). Mouse SULT1E1 was co-crystallized with PAP (desulfonated PAPS) followed by soaking with estradiol. This enabled the structural features important for both cofactor and substrate binding to be determined. Subsequently, the crystal structures of two other cytosolic sulfotransferases were determined, i.e., human dopamine/catecholamine sulfotransferase (SULT1A3; Refs. 110 and 111) and human hydroxysteroid sulfotransferase (SULT2A1; Ref. 112). Importantly, key structural elements are completely conserved and superimposable in all three enzymes. Furthermore, the structural features of the cofactor binding site, i.e., the PSB loop, as well as the 5'-phosphate and 3'-phosphate binding sites (cf. Fig. 6), are conserved in the three crystal structures. To date, however, the crystal structure of only the mouse estrogen sulfotransferase has been solved in the presence of substrate (109). Recently, the crystal structure of the human SULT1E1-PAPS complex was solved, which represents the first structure containing the active sulfonate donor for any sulfotransferase (113). The latter study has revealed a crucial reaction mechanism involving a 3'-phosphate-serine¹³⁷ interaction. This helps to regulate the action of lysine⁴⁷ in controlling the dissociation of the 5'-sulfate group form PAPS (113).

2. Mutagenesis.
In addition to crystallization, important structural information has also been garnered from studies utilizing site-directed mutagenesis to investigate a specific functionality. For instance, the structural basis for the high substrate specificity of SULT1E, which sulfonates the 3- hydroxyl group of phenolic steroids such as estrone and estradiol but not the 3-hydroxyl group of neutral steroids such as pregnenolone or DHEA (141), has been examined. It was determined that the 3ß-hydroxyl group of DHEA is excluded from the active site by steric hindrance of a tyrosine at position 81with the C-19 methyl group of DHEA creating a substrate gating phenomenon (142). In another case, the chirality exhibited by the two guinea pig SULT2A1 isoforms that demonstrate strict stereospecificity for either a 3- hydroxyl or a 3ß-hydroxyl group (143) was investigated. It was determined that the type of residue at position 51 has a significant ability to regulate this stereoselectivity, i.e., if the residue is an asparagine, -activity predominates, whereas if an isoleusine is in that position, ß-activity prevails (144). It is anticipated that solving the crystal structure of additional steroid sulfotransferases in the presence of their substrates will further elucidate underlying structural principles that determine substrate specificity.

Steroid sulfotransferases are generally homodimers in solution, and the structural elements responsible for the dimerization have been identified (145). A 10-residue segment near the carboxy terminus of SULT-proteins forms a hydrophobic zipper-like structure further enforced by ion pairs at both ends. This amino acid stretch, which includes a critical valine, is conserved as a KxxxTVxxxE motif in nearly all cytosolic sulfotransferases and appears to be the common protein-protein interaction motif mediating dimerization phenomena (145).

E. Biology
There continues to be a broad interest in exploring the potential biological role of steroid sulfonation in brain development and function (146, 147). This is particularly so for the sulfonates of DHEA (148, 149, 150, 151) and pregnenolone (152, 153, 154, 155). These latter studies are in addition to the well-established nongenomic actions of the sulfates of DHEA and pregnenolone on the GABA_A and N-methyl-D-asparate receptors (156). Importantly, steroid sulfotransferase activity and proteins have now been detected in brain tissue (157, 158, 159, 160).

In humans, DHEA, together with its sulfate, comprises the most abundant steroid in the circulation; however, the physiological significance of this phenomenon remains elusive. The plasma concentrations of DHEA and DHEA sulfate, which are produced by the zona reticularis of the adrenal cortex (161), undergo a dramatic age-related decline throughout adulthood from their postpubertal peak (162, 163). On the other hand, antiparallel to the decline of DHEA and DHEA sulfate, plasma cortisol levels show an increase (164). Chiefly as a result of this phenomenon, DHEA and DHEA sulfate have been implicated in a variety of physiological systems as well as pathophysiological disorders (165). There is a particular interest in these steroids in the aging process (165, 166). As a result, their use as potential therapeutic or counter-aging agents continues unabated, unfortunately, however, with conflicting results (165).

There also continues to be an interest in potential roles for steroid sulfotransferases in the induction and maintenance of endocrine-dependent cancers. This pertains principally to breast cancer (167, 168, 169, 170, 171) and carcinoma of the prostate gland (172). In theory, the formation and hydrolysis of steroid sulfates such as estrone sulfate in breast tumors or testosterone sulfate in prostate tumors could be an important mechanism regulating the availability of unconjugated steroids to interact with cognate nuclear receptors. Because only the unconjugated steroid hormone has growth-promoting activity, abnormal regulation of sulfotransferases has pathological implications. Furthermore, specific steroid hormone sulfotransferase transfection experiments have demonstrated effective reductions in cellular responses to physiological hormone concentrations (168, 172). Work in this area has led to the proposal that a potential target for antitumor therapy could be steroid sulfatase, the enzyme responsible for the hydrolysis of steroid sulfates (171). Although these various reports are of interest, there is no convincing evidence to support an etiological or pathophysiological role for any steroid sulfotransferase or sulfatase in endocrine-dependent tumorigenesis.

The biological effect produced by certain drugs and xenobiotics by targeting specific steroid sulfotransferases has been an interesting recent development (173, 174). By blocking the formation of a steroid hormone sulfate, thus increasing availability of the unconjugated form to interact with its cognate nuclear receptor, a xenobiotic can, in effect, exert an indirect hormonal action. Furthermore, this effect may not necessarily be associated with a significant change in the circulating steroid hormone level but may simply reflect a local expression in hormone-sensitive tissue.

A powerful technique for evaluating the biological significance of specific genes is to selectively inactivate the gene in question in an animal model. For example, targeted gene disruption has been used to evaluate estrogen sulfotransferase (175). It has been determined that disruption of the mouse SULT1E gene led to structural lesions in the adult male testis. Although knockout males are initially fertile and phenotypically normal, they eventually develop age-dependent Leydig cell hypertrophy/hyperplasia and seminiferous tubule damage. Additionally, older mice have reduced sperm motility and produce smaller litters compared with age-matched wild-type males (176). Disruption of the mouse SULT1E gene in female mice results in a significant reduction in fertility, although the nature of the defect is not known (176). It is hoped that this approach, which is important for evaluating the developmental and physiological significance of sulfoconjugation, will expand to involve other steroid sulfotransferase genes.

	V. Sulfonation and Thyroid Function

Top Abstract I. Introduction II. The Universal Sulfonate... III. Sulfotransferases IV. Sulfonation of... V. Sulfonation and Thyroid... VI. Sulfonation of... VII. Sulfonation of Peptide... VIII. Sulfonation of... IX. Summary and Future... References

 
A. General
The raison d’être of the thyroid gland is to produce the hormones T₄ and T₃, a process that involves the convergence of intrathyroidal iodine metabolism and protein synthesis. Thyroglobulin is iodinated at specific tyrosine residues followed by precise coupling steps, which subsequently lead to the formation of peptide-linked tri- and tetraiodothyronines. Eventually, proteolysis occurs with the release of T₄ and T₃ for secretion into the circulation while any residual iodine is recycled (177). Sulfonation plays an intimate role in thyroid physiology: it is important for both synthesis (178) and metabolism of thyroid hormones (179). In addition, TSH is subject to posttranslational modification by sulfonation (180, 181).

B. Sulfonation of thyroglobulin
Thyroglobulin is the major protein produced by the thyroid gland and the macromolecular precursor of thyroid hormones (182). During the course of synthesis and processing, thyroglobulin undergoes extensive posttranslational modification, including iodination (183), glycosylation (184), phosphorylation (185), and sulfonation (186). Modification of thyroglobulin by sulfonation involves both carbohydrate units and peptide chains. For instance, galactose-3-sulfate and N-acetylglucosamine-6-sulfate have been identified in the asparagine-linked complex of human thyroglobulin (187). Galactose-3-sulfate occurs primarily as a terminally linked residue, where it forms a novel-capping group. Additionally, tyrosine residues in the protein core of thyroglobulin are also subject to sulfonation (188). The fact that tyrosine sulfonation is a common posttranslational modification of thyroglobulin found throughout the vertebrate phylum suggests that it was acquired at an early stage in thyroid evolution (188). Interestingly, it has been demonstrated that TSH down-regulates thyroglobulin sulfonation, particularly on tyrosine residues (189).

The biological role of the sulfonated sugars in thyroglobulin (both the terminally capped galactose-3-sulfate as well as the more internally located N-acetylglucosamine-6- sulfate) remains uncertain. It has been speculated that sulfate groups linked to oligosaccharide chains or to tyrosine residues of thyroglobulin might be involved in protein-protein recognition and have implications for the proper folding of thyroglobulin (189). Furthermore, sulfated tyrosines in thyroglobulin may be involved in hormonogenesis, and the control of tyrosine sulfonation by TSH could be related to this process. The sulfate group could act to modify the electronic environment of the aromatic ring of tyrosine whereby it becomes more reactive for iodination and coupling; the sulfate group could be a recognition signal for thyroid peroxidase or the iodotyrosine donor (189). In a recent study (178), thyroid hormone synthesis was found to correlate with the sulfotyrosine content of thyroglobulin. Interestingly, a consensus sequence (Asp/Glu-Tyr) is similarly involved in thyroid hormone synthesis (190) and the sulfonation of tyrosine (191). Based on this observation, it has been proposed that thyroglobulin sulfotyrosines may act either as a signal for iodination or the coupling reaction (the latter is thought to be the more likely possibility) by inducing an interaction between thyroid peroxidase and thyroglobulin (178). Understanding the exact details of this complex system is a work in progress.

C. Sulfonation of thyroid hormones
T₄, the main secretory product of thyroid follicular cells, is converted in extrathyroidal tissues to T₃, the biologically active form of thyroid hormone (192). T₄ is activated by outer-ring deiodination (ORD) to T₃, and both T₄ and T₃ are inactivated by inner-ring deiodination (IRD) to rT₃ and diiodothyronine (DIT), respectively (Fig. 7). ORD and IRD are carried out by the enzyme type I iodothyronine deiodinase, an enzyme found primarily in the liver, kidney, and thyroid (193). Importantly, T₄ and T₃, as well as other iodothyronines, are subject to sulfoconjugation (194). Sulfonation of the phenolic hydroxyl group of T₄ and T₃ (Fig. 7) is carried out by a number of tissues including human liver (193). The sulfates of T₄ and T₃ are normal components of human serum, in which levels are subject to physiological and pathophysiological conditions (195, 196). T₃ sulfate is detectable in the fetal circulation, and its concentration increases with fetal age (197). Although T₃ sulfate does not bind to nuclear receptors and therefore lacks intrinsic biological activity (198), sulfonation of iodothyronines is nevertheless an important metabolic step in determining their disposal (179). For instance, the majority of normal T₃ disposal occurs via T₃ sulfate formation (199). It has long been known that sulfonation facilitates the deiodination of iodothyronines (200). Type I deiodination of T₃ sulfate in human liver homogenates occurs at a rate that is 30-fold higher as compared with unconjugated T₃, and HepG2 cells, which are deficient in T₃-sulfonating activity, are unable to deiodinate T₃ (201). It has been suggested that the function of sulfonation is to inactivate thyroid hormone so that iodine can be reused for thyroid hormone synthesis (202). This intriguing idea is supported by the demonstration that the sulfonation of T₄ and T₃, as well as other metabolites, strongly promotes hepatic deiodination (202). It is concluded that sulfonation simultaneously accelerates the IRD of both T₄ and T₃ (inactivation) and blocks the ORD of T₄ (activation) and thus represents a crucial step in the irreversible inactivation of thyroid hormones (203). On the other hand, if for some reason (e.g., propylthiouracil treatment) iodothyronine deiodinase activity is low, then T₃ sulfate becomes a reversible pathway in that biologically active unconjugated T₃ can be regenerated by the action of a tissue sulfatase (193). Akin to the sulfonation of thyroglobulin, the metabolism of iodothyronines by sulfoconjugation is an ever-evolving subject.

View larger version (14K): [in this window] [in a new window]   Figure 7. Stepwise deiodination of T4 by ORD to T3 and by IRD to rT3 and further by IRD of T3 and ORD of rT3 to 3,3'-DIT. Potential sulfonation acceptor sites are shown by (SO3-).

E. Sulfotransferases
The sulfotransferase enzymes that modify thyroglobulin, involving both the carbohydrate chains and the core peptide, are membrane-associated enzymes found in the Golgi complex. The polypeptide chain of thyroglobulin is synthesized in the endoplasmic reticulum, in which carbohydrate chain synthesis is also initiated. On formation of stable dimers, the nascent protein enters the Golgi system, where the carbohydrate units are completed and sulfonation takes place (177).

1. Enzymes sulfonating the carbohydrate chains of thyroglobulin.
Galactose 3-O-sulfotransferase activity involved in the formation of the galactose-3-sulfate capping groups present in the asparagine-linked oligosaccharides of thyroglobulin is located in the Golgi compartment (204). Four human galactose 3-O-sulfotransferases sharing approximately 40% identity have been cloned that carry out the sulfonation of different acceptor substrates (205). Of the four galactose 3-O-sulfotransferases, one was found to be highly expressed in the thyroid gland and is considered responsible for the formation of galactose-3-sulfate in ß14 linkage to N-acetylglucosamine attached to both N-glycans, and core 2-branched O-glycans synthesized in the thyroid (206). Although several 6-O-sulfotransferases, including N-acetylglucosamine 6-O-sulfotransferases, have been cloned (87), the sulfotransferase utilizing thyroglobulin (187) as an acceptor has not been identified.

2. Enzymes sulfonating tyrosine residues of thyroglobulin.
Protein tyrosine sulfonation is a widespread posttranslational modification of many secretory and membrane proteins (7, 8). Tyrosylprotein sulfotransferase is a 50,000- to 54,000-molecular-weight integral membrane protein of the trans-Golgi network that is found in essentially all tissues (207). Two distinct human tyrosylprotein sulfotransferase genes have been identified and the cDNAs cloned (106, 107, 208). The two tyrosylprotein sulfotransferases are 64% identical, with most of the variation between the two proteins found within the ends of the amino and carboxy termini (107). Although Northern analysis revealed that both genes were expressed in all the human tissues examined, expression by the thyroid gland has not been specifically investigated. Whether the two enzymes are functionally redundant or whether they might utilize preferred substrates is not known.

3. Enzymes sulfonating thyroid hormones.
In contrast to the Golgi-associated sulfotransferases that act on macromolecules, the enzymes that utilize iodothyronines as substrates are part of a large family of SULTs. Human SULTs with documented activity toward iodothyronines include SULT1A1 (thermostable phenol SULT; Refs. 209, 210, 211, 212, 213), SULT1A3 (thermolabile phenol SULT; Refs. 209 and 212), SULT1B1 (214, 215), SULT1E1 (estrogen SULT; Refs. 216 and 217), SULT1C1 (218), and SULT2A1 (hydroxysteroid SULT; Ref. 216). These multiple human SULT isoforms have overlapping specificities involving simple phenols, dopamine, estrogens, hydroxysteroids, as well as iodothyronines. A recent in vitro study (219) examined which SULT form makes a major contribution to the metabolism of T₃ by comparing five different overexpressed and purified human SULTs, i.e., SULT1A1, SULT1A3, SULT1B1, SULT1E1, and SULT2A1. Of the five enzymes, SULT1B1 demonstrated the lowest Michaelis-Menten constant (K_m) and the highest general rate constant (k_cat)/K_m for T₃; whereas SULT1A1 preferred the simple phenol p-nitrophenol, SULT1A3 preferred dopamine, SULT1E1 preferred estradiol-17ß, and SULT2A1 preferred DHEA (219). These results are consistent with the previous characterization of these SULT forms. Although the actual in vivo role of an individual SULT form in the metabolism of iodothyronines remains to be determined, this in vitro study clearly suggests that SULT1B1 is the principal SULT form involved in the metabolism of T₃. Furthermore, the content of SULT1B1 in human liver highly correlates with T₃-sulfonating activity (219), adding further support for SULT1B1 being the principal hepatic iodothyronine-sulfonating enzyme. SULT1B1 is expressed in the small intestine and colon, as well as in the liver; however, expression by thyroid tissue was not determined (215). SULT1C1, which was not included in the previous SULT comparison study, also sulfonates iodothyronines and is expressed in the thyroid gland, as well as in a number of other tissues (220). During fetal brain development, SULT activity with DIT as substrate correlated with SULT1A1 expression, suggesting that this SULT form is primarily responsible for the sulfonation of DIT (221). In summary, the involvement of a specific SULT form expressed in a specific tissue in the metabolism of iodothyronines remains to be fully elucidated.

	VI. Sulfonation of Catecholamines

Top Abstract I. Introduction II. The Universal Sulfonate... III. Sulfotransferases IV. Sulfonation of... V. Sulfonation and Thyroid... VI. Sulfonation of... VII. Sulfonation of Peptide... VIII. Sulfonation of... IX. Summary and Future... References

 
A. General
Conjugation represents a major metabolizing mechanism for catecholamines to the extent that approximately 84% of total epinephrine, 73% total of norepinephrine, and 97% of total dopamine circulate in a conjugated form (222, 223, 224). In man, the conjugated form of catecholamines is almost entirely sulfoconjugation, which is in contrast to rats, in which glucuronidation predominates (223). In addition, norepinephrine is methylated extraneuronally by the enzyme catechol O-methyltransferase to produce normetanephrine, and 77% of total circulating normetanephrine is also found in a sulfoconjugated form; again, this is in contrast to the rat, in which 63% of circulating normetanephrine is glucuronidated (225).

Free catecholamines exhibit a short plasma half-life of 1–3 min, in contrast to catecholamine sulfates, which have a plasma half-life of 3–4 h (226). Normotensive recumbent subjects exhibit an early (2300 h) nocturnal increase in the plasma concentration of the sulfates of dopamine, norepinephrine, and epinephrine (227); however, an explanation for this nocturnal rise has not been forthcoming. Catecholamines are sulfoconjugated predominantly at carbon-3 of the phenyl ring (Fig. 8), and the phenol sulfotransferase (SULT1A3) that carries out this reaction has the greatest affinity for dopamine, followed by norepinephrine and epinephrine (228). For plasma dopamine, which is more than 95% sulfonated, both the 3-O- and 4-O-sulfate isomers are present, with the 3-O-sulfate being in greater abundance by an order of magnitude (229).

View larger version (14K):