If you don't remember your password, you can reset it by entering your email address and clicking the Reset Password button. You will then receive an email that contains a secure link for resetting your password
If the address matches a valid account an email will be sent to __email__ with instructions for resetting your password
To whom correspondence should be addressed: Oklahoma Medical Research Foundation, 825 NE 13th St., Mailstop 45, Oklahoma City, OK 73104. Tel.: 405-271-7314; Fax: 405-271-7417;
Cardiovascular Biology Research Program, Oklahoma Medical Research Foundation, the Department of Medicine, University of Oklahoma Health Sciences Center, and the Oklahoma Center for Medical Glycobiology, Oklahoma City, Oklahoma 73104
* This minireview will be reprinted in the 2003 Minireview Compendium, which will be available in January, 2004. This work was supported by Grant HL63152 from the United States Public Health Service. The on-line version of this article (available at http://www.jbc.org) contains Supplemental Table I.
The abbreviations used are: TPST, tyrosylprotein sulfotransferase; PAPS, adenosine 3′-phosphate 5′-phosphosulfate; EST, expressed sequence tag data base; PSGL-1, P-selectin glycoprotein ligand-1; GPCR, G-proteincoupled receptor; HIV, human immunodeficiency virus; RANTES, regulated on activation normal T cell expressed and secreted; APS, adenosine phosphosulfate.
1The abbreviations used are: TPST, tyrosylprotein sulfotransferase; PAPS, adenosine 3′-phosphate 5′-phosphosulfate; EST, expressed sequence tag data base; PSGL-1, P-selectin glycoprotein ligand-1; GPCR, G-proteincoupled receptor; HIV, human immunodeficiency virus; RANTES, regulated on activation normal T cell expressed and secreted; APS, adenosine phosphosulfate.
EC 2.8.2.20). The enzyme was shown to catalyze the transfer of sulfate from the universal sulfate donor adenosine 3′-phosphate 5′-phosphosulfate (PAPS) to the hydroxyl group of a peptidyltyrosine residue to form a tyrosine O4-sulfate ester and 3′,5′-ADP (Fig. 1) (
) further showed that tyrosine O-sulfation occurred in the trans-Golgi network, and biochemical studies indicated that the enzyme was membrane-bound and that its active site was luminally oriented in highly enriched Golgi membranes.
Fig. 1The tyrosylprotein sulfotransferase reaction. TPSTs (EC 2.8.2.20) catalyze the transfer of sulfate from the universal sulfate donor PAPS to the hydroxyl group of a luminally oriented peptidyltyrosine residue to form a tyrosine O4-sulfate ester and 3′,5′-ADP.
) reported the purification of TPST from rat submandibular glands in 1997. Mouse and human cDNAs encoding TPST-1 were first isolated in our laboratory using amino acid sequences obtained from a purified rat liver TPST (
). Each TPST cDNA encodes a protein with type II transmembrane topology with a short N-terminal cytoplasmic domain, a single ≈17-residue transmembrane domain, and a luminal catalytic domain (Fig. 2). TPST-1 and -2 are of similar size (370 –377 residues), and each has two N-glycosylation sites and six conserved luminal cysteine residues. Two structural motifs found in cytosolic and membrane-bound sulfotransferases are conserved in TPST-1 and -2. In the known sulfotransferase crystal structures, these motifs are involved in binding of the 5′-and 3′-phosphate groups of the reaction product 3′,5′-ADP and are designated the 5′-PSB and 3′-PB motifs, respectively (
). The predicted amino acid sequences of human and mouse TPST-1 are ≈96% identical. Human and mouse TPST-2 have a similar degree of identity. Human and mouse TPST-1 are ≈65– 67% identical to human and mouse TPST-2, respectively. Multiple sequence alignments of TPSTs from various species show that the membrane-proximal portion of the luminal domain is poorly conserved. This ≈40-amino acid segment likely represents a “stem” region that may be dispensable for catalysis, analogous to that found in many glycosyltransferases. The human TPST1 and TPST2 genes are on 7q11.21 and 22q12.1, respectively, whereas the mouse Tpst1 (Mouse Genome Informatics accession number MGI:1298231) and Tpst2 (MGI:1309516) genes are both on chromosome 5, ∼18.5 Mb apart. There is no evidence for the existence of additional mouse or human TPST genes in genomic or expressed sequence tag (EST) data bases.
Fig. 2Domain structure of tyrosylprotein sulfotransferases. Human TPST-1 and -2 are type II transmembrane proteins of similar size (370 –377 residues). Each has a short 8-residue N-terminal cytoplasmic domain, a single ≈17-residue transmembrane domain (TM), a putative ≈40-residue stem region, a luminal catalytic domain, and two N-glycosylation sites (lancet). Four luminal cysteine residues that are conserved in TPSTs from all species are shown, as are the 5′-PSB and 3′-PB motifs that are involved in binding of the 5′- and 3′-phosphate groups of the reaction product 3′,5′-ADP, respectively.
Species and Tissue Distribution of Tyrosine O-Sulfation
Tyrosine-sulfated proteins and/or TPST activity have been described in many species throughout the plant and animal kingdoms, including the green algae, Volvox carteri, one of the earliest multicellular organisms (
). In vivo metabolic labeling studies in the rat have shown that [35S]sulfate is incorporated as tyrosine [35S]sulfate into many secretory and membrane proteins in all tissues examined (
Y. B. Ouyang and K. L. Moore, unpublished observations.
The many tissue sources of the several hundred human and mouse TPST-1 and -2 EST clones in the EST data base are also consistent with a widespread tissue and cellular distribution of TPST-1 and -2 transcripts. To date, no mammalian cell type or cell line has been described that does not express both transcripts. It is important to note that the distribution and relative abundance of the two TPST isoenzymes at the protein level in tissues or cells has not been examined due to the lack of isoenzyme-specific antibodies and substrates. Nevertheless, taken together, these data indicate that TPST-1 and -2 are broadly co-expressed in mammalian cells.
Searches of the EST data base reveals cDNAs encoding TPST-1 and/or TPST-2 orthologs in many other vertebrate (rat, dog, cow, pig, chicken, zebrafish, fugu, channel catfish, and African clawed frog) and invertebrate species (Caenorhabditis elegans, Drosophila melanogaster, Anopheles gambiae, Ciona intestinalis, Halocynthia roretzi, and Schistosoma japonicum).
We have sequenced several additional TPST cDNAs and deposited the sequences in GenBankTM, including D. melanogaster TPST (AY124548), H. roretzi TPST-2 (AY079190), Ictalurus punctatus TPST-2 (AF510735), Danio rerio TPST-1 (AF510736), and D. rerio TPST-2 (AF510737).
It is interesting to note that D. melanogaster has only a single TPST gene, unlike most other species, including C. elegans, which have two TPST genes. Tyrosine-sulfated proteins (
) have been described in several plant species, but no plant TPST orthologs have yet been identified, and none is apparent in the completed Arabidopsis thaliana genome. Furthermore, no tyrosine-sulfated proteins, TPST activity, or putative TPST orthologs have been described in prokaryotes or in yeast.
Substrate Specificity of Tyrosylprotein Sulfotransferases
Consistent with previous biochemical data and primary structural predictions, both TPST-1 and -2 expressed in CHO cells as N-terminal green fluorescent protein fusion proteins are localized in the Golgi as assessed by confocal microscopy.
D. Corbeil and W.B. Huttner, personal communication.
The Golgi localization and the luminal active site orientation of TPST-1 and -2 predict that tyrosine O-sulfation occurs only on proteins that transit the trans-Golgi network (Fig. 3). In keeping with this prediction, no example has been described that violates this rule. Table I (Supplemental Data) lists all tyrosine-sulfated proteins described to date. Secretory and transmembrane proteins of many different types and classes are represented with secreted proteins comprising the majority. This is consistent with pulse/chase and long term dual metabolic labeling studies of various cell lines that demonstrate that most protein-bound tyrosine sulfate is in secreted proteins (
). Based on these dual labeling experiments, it was estimated that, on average, 1 in 20 of the proteins secreted by HepG2 cells and 1 in 3 of those secreted by fibroblasts contain at least one tyrosine sulfate residue (
). These studies also indicate that membrane-bound proteins are as likely to be sulfated as secreted proteins. The relatively low amount of total protein-bound tyrosine sulfate in membrane proteins likely reflects their low abundance relative to secretory proteins.
Fig. 3Sulfate activation and tyrosine O-sulfation. Inorganic sulfate enters the cell by the action of one of several sulfate transporters (reviewed in Ref.
). These bifunctional enzymes contain a C-terminal ATP sulfurylase domain and an N-terminal adenosine phosphosulfate (APS) kinase domain. In the first step of sulfate activation, ATP and inorganic sulfate are converted to APS and pyrophosphate by ATP sulfurylase. APS is then channeled directly between the ATP sulfurylase and APS kinase active sites. In the second step catalyzed by the APS kinase domain, a second ATP is consumed to phosphorylate the 3′-hydroxyl of the ribose ring of APS to yield PAPS and ADP. PAPS is then transported into the Golgi lumen by a PAPS translocase that has been purified but not yet cloned (
). Once inside the Golgi lumen PAPS acts as the sulfate donor for TPSTs and all other carbohydrate sulfotransferases, and the sulfated products are either secreted or retained in the membrane of lysosomes, secretory vesicles, and/or the plasma membrane. TGN, trans-Golgi network.
The location of sulfation has been unambiguously defined in several proteins. Based on the amino acid sequences flanking known tyrosine O-sulfation sites, coupled with in vitro studies on the sulfation of various synthetic peptides, it is evident that there is no sequon for tyrosine O-sulfation per se. Although consensus features for tyrosine O-sulfation have been described, some proteins known to be tyrosine-sulfated do not fulfill these features (
). Nevertheless, it is clear that the dominant characteristic of known sulfation sites is that there are generally between 3 and 4 acidic amino acids within ±5 residues of the sulfotyrosine.
The SwissProt Group at the Swiss Institute of Bioinformatics has developed a software tool, called Sulfinator,
). Although the positive predictive value of this algorithm is uncertain, it can be used to estimate the number of tyrosine-sulfated proteins in a genomic scale dataset. Twenty-six percent of the 5421 mouse proteins in the SwissProt data base were predicted to have one or more tyrosine O-sulfation sites by the Sulfinator program. Data base entries were examined, and those proteins that lacked a predicted signal peptide, signal-anchor, and/or transmembrane domain were excluded, as were proteins in which the predicted sulfation site(s) is cytosolic in orientation, leaving 380 positive predictions which constitute 7% of the original 5421 proteins.
This suggests that as many as 2,100 mouse proteins might be tyrosine-sulfated, assuming that the mouse genome encodes 30,000 proteins. Thus, it is very likely that we are only beginning to appreciate the complexity of the substrate repertoire of these enzymes.
A great deal of interest in the field has recently focused on the role of tyrosine O-sulfation in G-protein-coupled receptor (GPCR) function after CCR5, a major HIV co-receptor, was shown to be tyrosine-sulfated (
). Site-directed mutagenesis and chlorate inhibition studies showed that sulfation of one or more tyrosine residues in the N-terminal extracellular domain of CCR5 are required for optimal binding of MIP-1α/CCL3, MIP-1β/CCL4, and RANTES/CCL5 and for optimal HIV co-receptor function. CCR5 mutants in which all four tyrosine residues in the N-terminal extracellular domain were changed to phenylalanine have ≈100-fold weaker affinity for MIP-1α/CCL3 and RANTES/CCL5. Other studies have shown that synthetic sulfotyrosine-containing peptides modeled on the N-terminal extracellular domain of CCR5, but not comparable phosphotyrosine-containing peptides, bound to the gp120-CD4 complex and inhibited HIV entry into cells (
). Likewise, mutagenesis studies indicate that sulfation of tyrosine residue(s) in the N-terminal extracellular domains of CXCR4, CCR2B, CX3CR1, C5a receptor, and the thyroid-stimulating hormone receptor is an important requirement for optimal binding of SDF-1α/CXCL12, MCP-1/CCL2, fractalkine/CX3CL1, C5a, and thyroid-stimulating hormone, respectively (
). In general, these studies have shown that ligand binding to mutated GPCRs or undersulfated wild-type GPCRs is between 5- and 100-fold weaker than binding to the native receptors. A comparison of the primary sequences of the known chemokine receptors shows that their N-terminal domains are highly acidic and contain one or more tyrosine residues. The fact that this is the dominant feature of known tyrosine sulfation sites suggests that many, and perhaps all, of the chemokine receptors may be sulfated. Chemokine receptors and their cognate ligands play crucial roles in innate and adaptive immunity, hematopoeisis, angiogenesis, tumor growth, and metastasis (
). Thus, the possibility that chemokine receptors as a class may require tyrosine sulfation for optimal chemokine binding has broad pathophysiological implications.
It is not known whether TPST-1 and -2 have distinct structural requirements for efficient sulfation of macromolecular substrates. However, in vitro studies using synthetic peptide acceptors indicate some differences in substrate preferences. We reported that peptides modeled on the sulfation sites of human C4 α chain and heparin cofactor II are more efficiently sulfated by TPST-1 than by TPST-2 (
). In addition, Seibert et al. studied sulfation of a peptide modeled on the N-terminal domain of human CCR5 that has four potential tyrosine O-sulfation sites (
). By determining the time course of appearance of various reaction products, they showed that each enzyme sulfated the peptide in a sequential fashion. Tyr14 and Tyr15 were sulfated first, followed by Tyr10 and then Tyr3. However, TPST-1 clearly preferred Tyr14 over Tyr15 as the initial sulfation site, whereas TPST-2 preferred Tyr15 over Tyr14. These studies also suggest that TPST-1 and -2 are not processive enzymes.
Regulation of Tyrosine O-Sulfation
Evidence that the TPST-1 or TPST-2 genes are subject to transcriptional regulation is very limited. Application of laminar shear stress to umbilical vein endothelial cells has been reported to inversely modulate TPST-1 and -2 transcript levels, but the changes are very modest (<2-fold) and of uncertain physiological significance (
). We have examined TPST-1 and -2 transcript levels in the liver and spleen of mice in response to intraperitoneal administration of lipopolysaccharide and in mouse lung and skin microvascular endothelial cells in response to tumor necrosis factor α. Over a 24-h period after stimulation, transcript levels were essentially unchanged.
Y. B. Ouyang, J. Deng, and K. L. Moore, unpublished observations.
It is possible that tyrosine O-sulfation might be modulated by the action of sulfatases. However, several lines of evidence argue that efficient mechanisms to desulfate tyrosine sulfate in intact proteins or tyrosine sulfate itself do not exist inside the cell or in the extracellular milieu. Dodgson et al. (
) reported that after intraperitoneal administration in rats, tyrosine [35S]sulfate was metabolized by deamidation, but the metabolites were excreted without loss of the sulfate ester. Furthermore, Jones et al. (
) reported that after intraperitoneal or intravenous injection of [35S]sulfate-labeled fibrinopeptide B in the rabbit, virtually all the injected radiolabel was recovered in the urine as free tyrosine [35S]sulfate or its deamidated metabolites. Tyrosine sulfate is also a normal constituent of human urine (≈28 mg/day) (
). Thus, even after degradative proteolysis, presumably in the lysosome, the sulfate ester of tyrosine O-sulfate is surprisingly stable. This is consistent with observations that tyrosine O-sulfate is a very poor substrate for mammalian sulfatases (
Several tyrosine-sulfated peptides and proteins have been purified from native sources, and the location and stoichiometry of sulfation have been determined. However, in most cases these proteins/peptides were isolated using bioassays to follow their purification and/or affinity chromatographic methods that depended on the presence of the sulfate esters, thus precluding an unbiased assessment of stoichiometry in vivo. Two exceptions are notable because they were clearly purified from an extracellular source, unlike most of the others, which were purified from tissues, and they were isolated using methods that were independent of the sulfation state of the protein. The activation peptide of human plasma-derived factor IX is stoichiometrically sulfated at Tyr155 (
). Human glycoprotein Ibα purified from human platelets has three tyrosine O-sulfation sites. Sulfation at Tyr278 and Tyr279 is near stoichiometric (>85%), whereas at Tyr276 sulfation is substoichiometric (≈50%) (
). Furthermore, cholecystokinin circulates in multiple stoichiometrically sulfated forms in human plasma each with a common C-terminal heptapeptide amide sequence -Tyr(SO3)-Met-Gly-Trp-Met-Asp-Phe-NH2 (
) showed that the circulating half-lives of [35S]sulfatelabeled fibrinopeptide B and intact fibrinogen in the rabbit are indistinguishable. These data indicate the absence of an efficient protein-tyrosine sulfatase activity in the extracellular space. Taken together these observations indicate that tyrosine O-sulfation is irreversible in vivo.
Role of Tyrosine O-Sulfation in Protein Function
To date, 62 tyrosine-sulfated proteins have been identified (Table I, Supplemental Data). For the majority of these, a role for sulfation in the function(s) of the proteins has not been described. Freiderrich et al. (
) reported that secretion of the D. melanogaster protein vitellogenin II from mouse fibroblasts was specifically inhibited when PAPS synthase activity was blocked using sodium chlorate or when the tyrosine O-sulfation site in vitellogenin II at Tyr172 was mutated to phenylalanine. However, inhibition of tyrosine O-sulfation does not affect surface expression or secretion of other tyrosine-sulfated proteins, including P-selectin glycoprotein ligand-1 (PSGL-1), glycoprotein Ibα, factor V, factor VIII, C4, and type III and V collagens. In other cases, sulfation clearly plays important roles in protein-protein interactions. This conclusion is most strongly supported by two examples in which tyrosine-sulfated proteins have been co-crystallized with ligand. Native hirudin is sulfated at Tyr63 and has a 10-fold tighter affinity for thrombin than unsulfated hirudin (
). In co-crystals of α-thrombin and hirugen (N-acetylhirudin-(53– 64), with sulfo-Tyr63), all three sulfato-oxygens of Tyr63 are involved in an extensive intermolecular hydrogen-bonded network with the anion binding exosite 1 of α-thrombin (
). In the co-crystal of the lectin-epidermal growth factor domain of P-selectin and a recombinant glycosulfopeptide mimetic of the N-terminal domain of PSGL-1, the sulfate groups at Tyr48 and Tyr51 are involved in direct protein-protein contacts (
). In other cases sulfation is important in optimal receptor-ligand interactions (e.g. chemokine/chemokine receptor binding), optimal proteolytic processing (e.g. gastrin processing), and proteolytic activation of extracellular proteins (e.g. factor V and VIII activation).
Until recently all of the accumulated data on the role of tyrosine O-sulfation in protein function has been derived from in vitro studies. In most cases in which a role for sulfation in the function of a protein has been defined, that function has been decreased in the absence of tyrosine O-sulfation but not absent. The only evidence that decreased tyrosine O-sulfation may have functional consequences in vivo comes from cases of mild to moderate hemophilia A that are due to missense mutations in the factor VIII gene that result in a Tyr1680 → Phe substitution in the von Willebrand factor binding site at the junction of the B and A3 domain (
). Sulfation of Tyr1680 in factor VIII is required for optimal binding to von Willebrand factor, which acts as a carrier protein for factor VIII in plasma, thereby increasing the circulating half-life of factor VIII in vivo (
) and attempts to generate Tpst1/Tpst2 double knock-outs are in progress. Northern analyses showed that disruption of one Tpst gene does not affect transcription of the other Tpst gene. Both Tpst1+/– and Tpst2+/– mice appear normal and when interbred yield litters of normal size with a Mendelian genetic distribution of the targeted mutation. Histological surveys of each knock-out have not revealed any abnormalities. A preliminary characterization of the Tpst1–/– mice revealed unexpected and pleiotropic effects of disruption of the Tpst1 gene, including effects on body weight, fecundity, and postnatal viability (
). Tpst1–/– mice appear healthy but have ≈5% lower average body weight than wild-type controls. We showed that fertility of Tpst1–/– males and females per se is normal. However, Tpst1–/– females have significantly smaller litters (≈4 pups/litter) than wild-type females (≈6 pups/litter) due to fetal death between 8.5 and 15.5 days postcoitum.
In contrast to Tpst1–/– mice, the growth of male and female Tpst2–/– mice was severely delayed. The maximum difference in body weight occurred at 4 weeks of age at which time the body weight of both male and female homozygotes was on average ≈20% below that of wild-type littermates. However, Tpst2–/– mice attained normal body weights at 10 weeks of age and otherwise appeared healthy to 12 months of age. A careful examination of the reproductive performance of Tpst2–/– mice revealed a severe defect in male but not female reproductive performance. In 20 different matings between Tpst2–/– males and either Tpst2–/– or Tpst2+/– females, only a single pup was sired by a Tpst2–/– male over a 2-month breeding period, despite the detection of vaginal plugs in most of the mating partners. Our preliminary analysis indicates that Tpst2–/– males have normal gonadal function, as assessed by normal testicular weights and serum follicle-stimulating hormone, luteinizing hormone, and testosterone levels. Epididymal sperm counts, sperm morphology, sperm motility, and testicular histology are normal. In addition, we observed no gross or histological abnormalities in the male reproductive tract. These data indicate that defective fertility of Tpst2–/– males may involve abnormalities in sperm transport, sperm capacitation, and/or fertilization per se. These possibilities are currently being investigated.
Many basic questions about tyrosine O-sulfation and the enzymes that catalyze its formation remain unanswered. Little is known about how TPST expression is regulated and how the two TPST isoenzymes differ with respect to substrate specificity. There is also not any information regarding the relative abundance of TPST-1 and -2 in cells. Targeted disruption of the Tpst1 and Tpst2 genes in mice result in distinct phenotypic effects. Currently we cannot directly link the phenotypic effects of TPST-1 or TPST-2 deficiency to defective sulfation of any known TPST substrate(s). Nevertheless, these data offer compelling evidence that the substrate repertoire of this enzyme system is substantially more complex than currently appreciated. Whether the distinct phenotypic effects observed reflect differences in substrate specificity, cellular distribution, and/or relative abundance of the two isoenzymes is not clear and will require further investigation. Nevertheless, these data demonstrate that the two TPST isoenzymes have distinct biological roles and underscore the importance of tyrosine O-sulfation in several physiological processes that had not been previously appreciated, including growth, regulation of body weight, and reproductive physiology.
The nucleotide sequence(s) reported in this paper has been submitted to the GenBank™/EBI Data Bank with accession number(s) AY124548, AY079190, AF510735, AF510736, and AF510737
86872-0&id=gr1.jpg){kind=link}
86872-0&id=gr2.jpg){kind=link}
86872-0&id=gr3.jpg){kind=link}