We investigate molecular mechanisms and biologic implications of sulfonation, a fundamental process in the biotransformation of endobiotics, drugs, and xenobiotics. Sulfonation, the transfer of an SO3-1 group from the universal donor molecule 3´-phosphoadenosine 5´-phosphosulfate (PAPS) to an acceptor molecule, is essential for normal growth and development as well as for maintenance of the internal milieu. Sulfonated macromolecules such as glycosaminoglycans and proteoglycans are involved in cell-surface and connective tissue structures and bone formation. Tyrosine sulfonation is a widespread post-translational modification of many secretory and membrane proteins. Sulfonation of specific saccharide moieties modifies glycoprotein hormones, creating unique structural motifs with important functional consequences. Sulfolipids are concentrated in the brain, peripheral nerves, and reproductive tissues. Sulfonation also plays an important role in modifying low–molecular weight compounds, such as catecholamines, iodothyronines, neuroendocrine peptides, and cholesterol along with its metabolites, oxysterols, bile acids, vitamin D, and steroids. By modulating the availability of biologically active hormones, sulfonation can influence biologic activity regardless of whether such compounds act in their unconjugated or sulfoconjugated form or via a genomic or nongenomic mechanism. Thus, sulfoconjugating enzymes, i.e., sulfotransferases, can play an essential role in specific physiologic systems and associated disorders.
Human hydroxysteroid sulfotransferases SULT2A1, SULT2B1a, and SULT2B1b
Strott, Fuda, Lee, Javitt
Enzymes sulfoconjugating neutral steroids and sterols constitute a family of sulfotransferases that are members of a super family of cytosolic sulfotransferases (SULT). Designated SULT2, the steroid/sterol family is further divided into two subfamilies: SULT2A1 and SULT2B1. Whereas the SULT2A1 subfamily in humans consists of a single form, the SULT2B1 subfamily comprises two isoforms (SULT2B1a and SULT2B1b) that result from an alternative exon 1 and differential splicing. Thus, the SULT2B1 isoforms have unique amino-termini but are otherwise identical. The SULT2A1 subfamily has broad substrate predilections; it acts on 3-alpha- and 3-beta-hydroxylated steroids, phenolic steroids, and bile acids as well as on a variety of xenobiotics but does not use cholesterol as a substrate to any significant degree. On the other hand, the isoforms of the SULT2B1 subfamily have narrower substrate preferences that are confined to selective 3-beta-hydroxylated compounds; the enzymes do not use 3-alpha-hydroxylated steroids, phenolic steroids, or bile acids as substrates. For example, SULT2B1a avidly sulfonates pregnenolone, dehydroepiandrosterone to a lesser degree, and cholesterol essentially not at all. Conversely, SULT2B1b functions as the physiologic cholesterol sulfotransferase. The fact that the SULT2A1 and SULT2B1 isozymes are differentially expressed and display dissimilar substrate propensities strongly suggests that the isozymes play distinct biologic roles. Interestingly, the SULT2B1 isoforms are structurally different from SULT2A1 and all other known cognate cytosolic sulfotransferases. For example, the SULT2B1 isoforms have extended amino- and carboxy-terminal ends that have functional significance, the unique amino-terminus of the SULT2B1b isoform is essential for catalytic activity, and the common carboxy-terminal region of the two isoforms is enriched in prolines as well as in glutamate, serine, and threonine, suggesting polypeptide sequences that target proteins for rapid degradation.
Strott CA. Hydroxysteroid sulfotransferases SULT2A1, SULT2B1a and SULT2B1b. In: Coughtrie MWH, Pacifici GM, eds. Human Cytosolic Sulfotransferases. London: Taylor & Francis, 2005;231-251.
Human cholesterol sulfotransferase and cholesterol sulfate
Hanyu, Fuda, Lee, Strott; in collaboration with Higashi, Ikegawa
Cholesterol sulfate is abundant in human plasma; it is quantitatively more significant than other known sterol sulfates and is implicated in several biologic systems. It is a component of cell membranes, where it has a stabilizing effect, e.g., protecting erythrocytes from osmotic lysis; it can regulate serine proteases such as those involved in blood clotting, fibrinolysis, and epidermal cell adhesion; and it has been implicated in adrenal and gonadal steroid synthesis, regulation of cholesterol and fatty acid synthesis, sperm capacitation, platelet aggregation, and, because it can regulate selective protein kinase C isoforms and modulate the specificity of phosphatidylinositol 3-kinase, in signal transduction. In keratinocyte differentiation, cholesterol sulfate can induce genes that encode key components of barrier formation. Along with indications of cholesterol sulfate’s emerging role as a regulatory molecule, the fact that cholesterol sulfate circulates in blood suggests that it acts as a hormone.
Cholesterol sulfotransferase (SULT2B1b) is the only SULT2 enzyme expressed in normal human skin and normal human epidermal keratinocytes (NHEK). In primary cultures of NHEK, Ca2+-induced of expression of SULT2B1b mRNA, paralleled by increases in SULT2B1b protein and enzymatic activity, is consistent with the involvement of cholesterol sulfate in keratinocyte development. Furthermore, adding cholesterol sulfate to NHEK cultures under low-calcium conditions induces expression of the gene encoding involucrin, an important barrier protein. However, subsequent studies revealed that the time-course and patterns of SULT2B1b mRNA expression are similar to those of involucrin, filaggrin, transglutaminase I, and the nuclear receptor RORalpha and thus inconsistent with immunocytochemical analyses of SULT2B1b expression. These studies showed that expression is confined to the granular layer of the epidermis and thus suggest that SULT2B1b is a late marker of keratinocyte differentiation (an expression pattern similar that of the barrier protein, filaggrin, a late marker of differentiation). On the other hand, cells express involucrin, another barrier protein, and transglutaminase I, a key protein cross-linking enzyme, in the immediate suprabasal region or spinous layer of the epidermis; such cells are the earliest to appear in the pathway toward keratinocyte differentiation and thus are early markers of differentiation. We are employing RNAi methodology to resolve this apparent paradox and define cholesterol sulfate’s role in regulating the expression of specific barrier proteins.
Confinement of cholesterol sulfotransferase to the granular layer of the living epidermis, which lies beneath the stratum corneum and appears to contain the highest content of cholesterol sulfate, strongly suggests that the sulfolipid’s principal function is carried out primarily in the region of the granular-stratum corneum junction. The epidermis is a perpetually renewing tissue in which keratinocytes arise from stem cells in the basal layer and move through a series of cellular differentiation events until, as dead squames, they are sloughed off from the outer stratum corneum (desquamation). Cholesterol sulfate can retard desquamation by inhibiting serine proteases in the stratum corneum that are responsible for the dissolution of cell adhesion structures known as desmosomes and can influence cell cohesiveness by affecting the stability of corneocyte lipid bilayers. We posit that the inhibitory influence of cholesterol sulfate on normal desquamation would be subsequently removed by the presence of a steroid sulfate sulfohydrolase in the outer stratum corneum, thus allowing normal desquamation to occur. Our hypothesis is supported by markedly elevated cholesterol sulfate levels in the stratum corneum of patients with recessive X-linked ichthyosis, a scaling disorder resulting from deficient steroid sulfatase activity and consequent abnormal desquamation. During Ca2-induction of keratinocyte differentiation, most of the cholesterol sulfate is found in the cellular membrane fraction. To determine cell type and membrane localization of cholesterol sulfate, we are developing an antibody to cholesterol sulfate in collaboration with Shigeo Ikegawa.
Cholesterol sulfate can also play a significant role in platelets by potentiating arachidonic acid–, ADP-, and thrombin-induced platelet aggregation as well as serotonin secretion. Platelets contain SULT2B1b, suggesting that the sulfate is produced locally. As in skin, SULT2B1b is the only SULT2 enzyme expressed by these discoid anucleate particles, which nonetheless engage in protein synthesis. Human platelets and plasma lipoproteins interact and are intimately involved in the pathogenesis of atherosclerosis, thrombosis, and coronary artery disease. HDL, specifically the apolipoprotein A-I component, stabilizes SULT2B1b mRNA in platelets, and we observed a direct correlation between platelet SULT2B1b mRNA and protein levels in the presence or absence of lipoprotein as reflected in enzymatic activity and cholesterol sulfate production.
Higashi Y, Fuda H, Yanai H, Lee Y, Fukushige T, Kanazaki T, Strott CA. Expression of cholesterol sulfotransferase (SULT2B1b) in human skin and primary cultures of human epidermal keratinocytes. J Invest Dermatol 2004;122:1207-1213.
Yanai H, Javitt NB, Higashi Y, Fuda H, Strott CA. Expression of cholesterol sulfotransferase (SULT2B1b) in human platelets. Circulation 2004;109:92-96.
The role of RORalpha in cholesterol sulfate action
Hanyu, Fuda, Strott; in collaboration with Gorshkova, Higashi, Schuck
The retinoic acid–related orphan receptor alpha (RORalpha) is a member of the subfamily 1 of nuclear hormone receptors. Recent structural and functional studies have led to the hypothesis that either cholesterol or a cholesterol derivative is the natural ligand of RORalpha. Others have now solved the x-ray crystal structure of the ligand-binding domain of RORalpha in complex with cholesterol sulfate following a ligand exchange. In contrast to the 3-hydroxyl of cholesterol, the 3-O-sulfate group makes additional direct hydrogen bonds with three residues of the RORalpha ligand-binding domain. A comparison of the complex with cholesterol revealed that seven well-ordered water molecules are displaced and that the ligand is slightly shifted toward the hydrophilic part of the ligand-binding pocket, which is ideally suited for interactions with a sulfate group. The additional ligand-protein interactions result in a higher affinity for cholesterol sulfate than for cholesterol, as shown by mass spectrometry analysis under native conditions and differential scanning calorimetry. Moreover, mutational studies have shown that the higher binding affinity of cholesterol sulfate translates into increased transcriptional activity of RORalpha. These findings suggest that, in part, the molecular action ascribed to cholesterol sulfate involves interactions with the RORalpha nuclear receptor. Importantly, RORalpha is highly expressed in primary cultures of NHEK. Furthermore, immunohistochemical analysis of normal human skin reveals nuclear staining for RORalpha in the outer granular layer of the epidermis, a pattern similar to that of SULT2B1b. We have now cloned, overexpressed, and purified human RORalpha for use in binding experiments to determine its cholesterol sulfate binding parameters.
Transcriptional regulation of human SULT2B1b expression
Lee, Luu, Chan, Fuda, Strott; in collaboration with Higashi
Studies involving RT-PCR reveal that human SULT2B1a and SULT2B1b are differentially expressed. Real-time PCR demonstrates that expression of SULT2B1b is high in the prostate, placenta, and skin; weak in the kidney, lung, small intestine, and colon; and nonexistent or almost nonexistent in other tissues; however, nothing is known about regulation of the two human SULT2B1 isoforms. We focused on the expression of SUT2B1b in NHEK as well as in immortalized but highly differentiated human keratinocytes (HaCaT cells).
The gene for human SULT2B1 contains neither a canonical TATAAA nor a CCAAT motif in the upstream region flanking exon 1B that encodes for the SULT2B1b isoform; it also lacks an initiator motif, an element overlapping the transcription start site (TSS) of many TATA-lacking and TATA-containing promoters. Using primer-extension, RNase protection, and the RNA ligase–mediated rapid amplification of cDNA ends (RACE), we found mutliple TSSs for SULT2B1 expression of the SULT2B1b isoform. While some TATA-less promoters direct transcription initiation from a specific nucleotide, others do so from several sites. Given the variability in potential start sites, we found that the 5´-end of the SULT2B1b mRNA extends beyond nt -161 and may extend to nt -174 upstream of the ATG initiation codon. Thus, we located the TSS for the SULT2B1b isoform between nt -161 and nt -174 upstream of the translation initiation codon.
Many TATA-less promoters are characterized by several GC boxes that bind to the Sp1 transcription activator and thus play a central role in the assembly of the transcription complex of the promoters. Regulation of the SULT2B1 gene expressing the SULT2B1b isoform is at least partly influenced by the Sp1 family of transcription factors, i.e., the area upstream of the coding region of SULT2B1b contains several GC/GT boxes, while mutational analyses indicated involvement of specific GC/GT boxes located between nt -215 and nt -221 and between nt -127 and nt -136 in transcriptional regulation. Furthermore, progressive shortening of the 5´-flanking region yielded results entirely consistent with the mutational experiments. Locating the TSS at nt -161 places the two regulatory Sp1 elements above and below the TSS, i.e., at 23 nt upstream of the TSS and 47 nt downstream of the TSS, positioning the latter within the 5´-UTR. Transcriptional regulation within the 5´-UTR is well recognized. Promoter activity within exon 1 represents a positive internal regulatory sequence, an effect commonly attributable to the presence of functional Sp1 binding elements.
We obtained further support for Sp1 regulation of the human SULT2B1 gene when we found that nuclear extracts from HaCaT cells expressing SULT2B1b contain proteins that bind to Sp1-containg probes as well as to Sp1 and Sp2 proteins. In addition, co-transfection of HaCaT cells with Sp1 and/or Sp2 expression vectors produced dose-dependent increases in promoter activity, although transcriptional activation by Sp2 was less potent than that produced by Sp1. Interestingly, co-expression of Sp1 and Sp2 in suboptimal amounts produced a synergistic effect. Use of a histone deacetylase (HDAC) inhibitor resulted in dramatic augmentation of reporter gene activity induced by Sp1, an effect not seen with Sp2. Furthermore, the synergism produced by Sp1 and Sp2 in the absence of trichostatin A (TSA) treatment was not demonstrable in the presence of TSA, i.e., the level of transcriptional activation with Sp1 and Sp2 in the presence of TSA was no greater than that occurring with Sp1 alone plus TSA. We thus concluded that, whereas Sp1 functions as a transactivator of the SULT2B1 gene regulating expression of the SULT2B1b isoform, the role of Sp2 is less clear despite Sp2’s apparent ability to enhance the effect of Sp1 in experiments that did not involve HDAC inhibition. Stimulation of promoter activity with TSA alone in the absence of Sp1 co-transfection might be partly attributable to the presence of endogenous Sp1 in HaCaT cells. The effect of HDAC inhibition was lost with mutation of the regulatory Sp1-binding sites, suggesting that Sp1 stimulated SULT2B1b promoter activity by inducing histone acetylation. In addition, it is possible that TSA activation of promoter activity is mediated by stabilizing acetylation of Sp1 itself, which in turn could recruit other factors to the active site.
Lee YC, Higashi Y, Luu C, Shimizu C, Strott CA. Sp1 elements in SULT2B1b promoter and 5´-untranslated region of mRNA: Sp1/Sp2 induction and augmentation by histone deacetylase inhibition. FEBS Lett 2005;579:3639-3645.
Cloning, characterization, and expression of rat SULT2B1 isoforms; analysis of rat SULT2B1 gene structure
Kojitani, Fuda, Hanyu, Strott
The SULT2B1 gene is unique among the steroid/sterol sulfotransferase genes in that it encodes two isoforms as a result of an alternative exon I. The two isoforms are thus identical except for their amino-termini. In humans and mice, a portion of exon IA encoding an amino acid sequence common to both isoforms is fused with exon IB to complete the SULT2B1b mRNA, which is achieved by differential splicing. The structure of the rat SULT2B1 gene demonstrates an interesting and unexpected variation on this theme, particularly given that rat and mouse species are more closely related than human and mouse species. In the rat SULT2B1 gene, exon IA is upstream of exon IB, which is the reverse of the situation in human and mouse genes, and that portion of exon IA encoding the unique amino-terminus of the SULT2B1a isoform is relocated. The part of exon IA encoding the common amino acid sequence of the two isoforms remains in the same relative gene position as in the human and mouse genes and becomes exon II. As a result, the rat does not require differential splicing, which produces the SULT2B1b mRNA in human and mouse species. The mechanisms underlying this precise exonic cutting and rearrangement are not yet understood.
Structure/function relationships of SULT2B1 isoforms in the three species reveal consistencies as well as differences. For instance, the rat and human SULT2B1a isoform does not sulfonate cholesterol to any significant degree, whereas the mouse SULT2B1a isoform sulfonates cholesterol more efficiently than either pregnenolone or DHEA while the mouse, rat, and human SULT2B1a isoform sulfonates pregnenolone much more vigorously than DHEA and cholesterol. In contrast, the SULT2B1b isoform in all three species sulfonates cholesterol with the highest efficiency and thus represents the physiologic cholesterol sulfotransferase. Interestingly, the unique amino-termini of the SULT2B1b isoform in the three species are similar in size and highly homologous. Based on deletion studies of the human SULT2B1 isoforms, the SULT2B1b amino-terminus, encoded by exon IB, is essential for cholesterol catalysis. While the amino-termini of the SULT2B1b protein of rat, mouse, and human are similar in size, the amino-termini of rat and mouse SULT2B1a are markedly longer than those of human SULT2B1a. By contrast, the functional significance of the unique amino-terminus of the SULT2B1a isoform is not known. Unlike the case of the SULT2B1b isoform, removal of the unique amino-termus of human SULT2B1a does not affect the enzyme’s ability to sulfonate pregnenolone.
We have also observed consistencies and differences in the tissue expression of mRNAs for the SULT2B1 isoforms in rat, mouse, and human species. In all three species, skin is the dominant tissue for expression of the SULT2B1b isoform, consistent with the role of cholesterol sulfate in epidermal differentiation and development. Human placenta and prostate significantly express SULT2B1b, whereas the rat placenta and prostate express it, respectively, barely and not at all; however, modest expression of SULT2B1b occurs in the mouse prostate. After skin, the next tissue of importance in SULT2B1b expression in the rat and mouse is the intestine, whereas the human intestine appears to express SULT2B1b weakly. The third tissue in humans and the rat that modestly expresses SULT2B1b is the kidney, whereas SULT2B1b is poorly expressed in the mouse kidney. The kidneys of male and female rats contain high concentrations of cholesterol sulfate, but its concentration in the rat prostate is low, a finding consistent with the level of SULT2B1b mRNA expression in rat kidney and prostate tissues. In the rat, the SULT2B1a isoform is expressed only in the brain and testis; in the mouse, almost exclusively in the brain. The adult human brain, however, does not express either SULT2B1 isoform (standard RT-PCR suggested that human fetal brain expresses SULT2B1a). Given that pregnenolone sulfate functions as a neurosteroid (at least in the rodent brain), we are interested in the expression of SULT2B1a in the brain of the rat and mouse and are now investigating further the enzyme’s expression in the brain by in situ hybridization.
Kohjitani A, Fuda H, Hanyu O, Strott CA. Cloning, characterization and tissue expression of rat SULT2B1a and SULT2B1b steroid/sterol sulfotransferase isoforms: divergence of the rat SULT2B1 gene structure from orthologous human and mouse genes. Gene 2005 [Epub ahead of print].
Role of SULT2B1b in protecting against age-related macular degeneration
Javitt, Fuda, Strott; in collaboration with Rodriguez
Age-related macular degeneration, the leading cause of blindness in elderly individuals, is a complex disease that involves the aging process as well as genetics and environmental factors. The accumulation of cholesterol in Bruch’s membrane as a process of aging as well as the epidemiologic association with atherosclerosis suggests related mechanisms. In atherosclerosis, the accumulation of low-density lipoproteins (LDLs) in arteries and subsequent oxidation and ingestion of LDLs by macrophages is believed to be critical in the formation of atherosclerotic plaques. The internalization of oxidized LDL (oxLDL) by macrophages leads to foam cell formation, a process thought to be one of the principal causes of atherosclerosis. The cytotoxicity of oxLDL has also been reported in aortic endothelial cells and retinal pigment epithelium (RPE) cells. The oxidation of the esterified and unesterified cholesterol within LDL particles generates a series of cholesterol oxides known as oxysterols that have potent pharmacological activities, including induction of apoptosis and necrosis. Using ARPE19 rat retinal cells in culture, we showed that oxLDL increased cytotoxicity with prolonged oxidation. Analysis of oxLDL demonstrates predominance of 7-oxygenerated products, including 7alpha-/7beta-hydroxycholesterol and 7-ketocholesterol (7kCh). The addition of these oxysterols to ARPE19 cells in free form showed 7kCH to be the most cytotoxic; prolonged oxidation of LDL increases the levels of 7kCh. Interestingly, SULT2B1b, but neither SULT2B1a nor SULT2A1, is expressed in retinal ARPE19 cells as well as in the normal monkey retina, and, importantly, SULT2B1b actively sulfonates 7kCh. Furthermore, free 7kCh is cytotoxic when added to cultures of ARPE19 cells, whereas the sulfoconjugation of 7kCh protects cells from the adverse effects of unesterified 7kCh.
COLLABORATORS
Inna Gorshkova, PhD, Division of Bioengineering and Physical Science, ORS, NIH, Bethesda, MD
Yuko Higashi, MD, PhD, Kagoshima University Graduate School of Medical and Dental Sciences, Kagoshima, Japan
Shigeo Ikegawa, PhD, Faculty of Pharmaceutical Sciences, Kinki University, Osaka, Japan
Ignacio R. Rodriguez, MD, PhD, Laboratory of Retinal Cell and Molecular Biology, NEI, Bethesda, MD
Peter Schuck, PhD, Division of Bioengineering and Physical Science, ORS, NIH, Bethesda, MD
For further information, contact chastro@mail.nih.gov.