THE BIOTRANSFORMATION
OF ENDOBIOTICS BY SULFONATION
Young C. Lee, Ph.D., Staff Scientist
Hirotoshi Fuda, Ph.D., Postdoctoral Fellow
Yuko Higashi, M.D., Ph.D., Postdoctoral Fellow
Chikara Shimizu, M.D., Ph.D., Postdoctoral Fellow
Hidekatsu Yanei, M.D., Ph.D., Postdoctoral Fellow
Norman B. Javitt, M.D., Ph.D., Guest Researcher
Adriana Burgos, HHMI Scholar
PAPS Synthase Characterization and Expression
Fuda, Lee, Shimizu, Strott
Sulfonation cannot occur in the absence of the universal sulfonate donor molecule 3'-phosphoadenosine 5'-phosphosulfate (PAPS). The production of PAPS from ATP and inorganic sulfate is carried out by a bifunctional enzyme, the product of two separate genes, i.e., PAPS synthase 1 and PAPS synthase 2. In addition, as a result of differential splicing, PAPS synthase 2 occurs as two subtypes, creating a total of three PAPS synthase isoforms. We have cloned and characterized both gene products in humans as well as in guinea pigs and rabbits. It is not clear why two genes regulate PAPS formation; furthermore, the biochemical and physiological relationships of the PAPS synthase isoforms remain poorly understood. We have examined differences in the genes’ amino acid composition and whether the differences in composition might be associated with functional differences. All three isoforms demonstrate Michaelis/Menten kinetics, but the PAPS synthase 2 isoforms have a greater catalytic efficiency (kcat/Km) and are 15 to 20 times more active than PAPS synthase 1. Multiplex RT-PCR reveals that PAPS synthase1 is expressed in essentially all human tissues, suggesting that the gene is constitutively expressed. In contrast, PAPS synthase 2 is expressed in a tissue-specific manner; moreover, the PAPS synthase 2 subtypes are differentially expressed, suggesting tissue-specific regulation of the gene as well as differential splicing.
Human PAPS synthase 2 was discovered during a search for the genetic basis of a developmental abnormality causing a form of spondyloepimeta-physeal dysplasia that presents with a skeletal phenotype involving the spine and long bones. Initially, 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 co-expression of PAPS synthase 1 in cartilaginous tissue. In fact, as we have been able to demonstrate, PAPS synthase 1 is more vigorously expressed than PAPS synthase 2 in cartilage from human adults. The pathology in human spondyloepimetaphyseal dysplasia is attributable to undersulfation of extracellular matrix proteoglycans. We believe that this apparent enigma has been resolved by demonstrating that, in the cartilaginous growth plate of guinea pig long bones, PAPS synthase 2 is indeed the predominantly expressed isoform, whereas PAPS synthase 1 is poorly expressed. Such appears to be a general phenomenon in immature animals, i.e., cartilage obtained from non-growth plate sites (e.g., sternum and rib cage) also more vigorously expresses PAPS synthase 2 in contrast to the situation with mature animals, which, as the human adults, more prominently express PAPS synthase 1 in their cartilage. It is highly likely that this finding involving immature guinea pigs’ cartilage as an animal model will also hold true for human development where tissues are not readily available for examination.
PAPS Synthase Transcriptional Regulation
Shimizu, Fuda, Lee, Strott
The gene for PAPS synthase 1 contains neither a TATAAA nor a CCAAT box. The RNA ligase–mediated rapid amplification of cDNA ends (RLM-RACE) method, a recently developed variation of the 5'RACE procedure, revealed the presence of two possible transcription start sites, as confirmed by RT-PCR experiments that used gene-specific primers. We determined the proximal promoter region and identified multiple GC/GT boxes. Two of the elements upstream of the proximal promoter region and two downstream GC/GT boxes nearest the transcription start site significantly influence promoter activity. Nuclear extracts of SW13 cells, which highly express PAPS synthase 1, contain proteins that bind to probes possessing specific GC/GT boxes. The gene for PAPS synthase 1 appears to be under the influence of the Sp1 family of transcription factors, i.e., supershift analysis confirmed the presence of Sp1, Sp2, and Sp3 proteins in the nuclear extract. Co-transfection experiments using SL2 cells yielded additional support for the involvement of Sp1 in transcriptional regulation of the PAPS synthase 1 gene.
As with the gene for PAPS synthase 1, the promoter for PAPS synthase 2 contains neither a TATAAA nor a CCAAT box, although a consensus initiator motif is present downstream of the cap site. We determined two transcription initiation sites similar to PAPS synthase 1 for PAPS synthase 2. We confirmed the putative start sites by RT-PCR experiments that used gene-specific primers. The proximal promoter for HepG2, JEG3, and SW13 cell lines is similarly located between bp -84 and bp -124 upstream of the putative transcription start site. The region contains two GC/GT boxes that are essential for full promoter activity, as indicated by deletion analysis and further supported by mutagenesis. A nuclear extract of SW13 cells, which highly express PAPS synthase 2 as well as PAPS synthase 1, contains proteins that bind to probes possessing promoter-specific GC/GT boxes; furthermore, similar to the case with PAPS synthase 1, supershift analysis confirmed the presence of Sp1, Sp2, and Sp3 proteins in the nuclear extract, while co-transfection experi-ments using SL2 cells yielded additional support for involvement of Sp1 in transcriptional regulation of the PAPS synthase 2 gene.
Kinetic Analysis and Differential Expression of Human Hydroxysteroid Sulfotransferase SULT2B1 Isoforms
Fuda, Lee, Higashi, Javitt, Strott
Due to the presence of an alternative exon 1, the human hydroxysteroid sulfotransferase gene SULT2B1 encodes two subtypes that structurally differ only at their amino termini. The influence of the unique amino termini on substrate specificity, however, has remained unclear. Interestingly, kinetic data demonstrate that the SULT2B1 isoforms and the prototypical SULT2A1 isozyme exhibit distinct substrate preferences. For instance, SULT2B1a avidly sulfonates pregnenolone but not cholesterol, whereas SULT2B1b selectively sulfonates cholesterol, suggesting that the latter isoform is the true human cholesterol sulfotransferase. Notably, in contrast to SULT2A1, which is regarded as a dehydroepiandrosterone (DHEA) sulfotransferase, neither SULT2B1 isoform efficiently sulfonates DHEA. These distinct substrate predilections, along with their differential expression patterns, strongly suggest that the SULT2 isozymes have distinct roles. Cholesterol sulfate is widely distributed and, in human plasma, is the predominant steroid sulfate. Because sulfonation of cholesterol is crucial to keratinocyte differentiation as well as to the development of the barrier and thus to normal skin development, the finding that SULT2B1b functions as a cholesterol sulfotransferase and is the predominant steroid/sterol sulfotransferase expressed in human skin has proved significant. In contrast, the prototypical steroid sulfotransferase SULT2A1 does not sulfonate cholesterol and is not expressed in human skin.
Structure/Function Analysis of the Human Hydroxysteroid Sulfotransferase SULT2B1 Isoforms
Fuda, Shimizu, Lee, Javitt, Strott
From a structural point of view, compared with the SULT2A1 isozyme, the outstanding feature of the SULT2B1 isoforms as well as other of cloned steroid and cognate cytosolic sulfotrans-ferases is their extended amino- and carboxy-terminal ends. All previously cloned members of the mammalian cytosolic sulfotransferase superfamily, i.e., estrogen and phenol sulfotrans-ferases as well as hydroxysteroid sulfotrans-ferases, are composed of 282 to 295 amino acids, whereas SULT2B1a and SULT2B1b consist of 350 and 365 amino acids, respectively. Nonetheless, the SULT2A1 and SULT2B1 isozymes exhibit a significant structural similarity in their core regions, most notably among the amino acid residues important in protein-PAPS co-factor interaction of cytosolic sulfotransferases, which are completely conserved.
The functional significance of the extended carboxyterminal end of the SULT2B1 isoforms is not known. One speculation is that this region, which is enriched in prolines, may play a role in protein-protein interactions. It is notable that the relatively long carboxyterminal extension (53 amino acids) common to both SULT2B1 isoforms can be removed without causing a significant change in the catalytic behavior of either isoform. On the other hand, removal of the unique amino-terminal ends, which distinguish the SULT2B1 isoforms, produces interesting results. Removal of the unique 23 amino-terminal residues of SULTT2B1b abolishes catalytic activity, whereas removal of the unique eight amino-terminal residues of SULT2B1a does not; indeed, the first 18 amino acids of the SULT2B1b amino-terminal region are not essential for catalytic activity, although the five-amino acid segment between residues 18 and 24 from the amino terminus of SULT2B1b is essential for catalytic activity. Within this five-amino acid DISEI sequence of SULT2B1b, it is the isoleucines that are critical for catalysis. Moreover, amino acid substitutions show that only a hydrophobic amino acid with a side chain of optimal size can function effectively at these positions, i.e., alanine cannot fulfill this function, whereas leucine and isoleucine work equally well; furthermore, methionine, although less effective than isoleucine, is also able to sustain significant enzymatic activity. Conversely, charged and polar amino acid substitutions at either of the isoleucine positions are inactive.
The human SULT2B1 gene consists of an exon 1B, an exon 1A, and exons 2 through 6: the SULT2B1a isoform is encoded by exons 1A and 2 through 6; the SULT2B1b isoform is encoded by exon 1B, the final 143 nucleotides of exon 1A plus exons 2 through 6. Exon 1B is composed of a 5'UTR consisting of 129 nucleotides and the coding region for the first 23 amino acids of SULT2B1b, which represents the entire amino-terminal region that is unique to this isoform. Exon 1A, on the other hand, in addition to a 179-nucleotide 5’UTR, encodes the first 56 amino acids of SULT2B1a, of which only the first eight amino acids are unique to this isoform. Thus, when exon 1B of the human SULT2B1 gene is transcribed, cholesterol sulfotransferase is synthesized, whereas when exon 1A is transcribed, pregnenolone sulfotransferase is produced.
For several reasons, the most suitable organ systems in which to consider the physiological implications of the differential expression of the human SULT2B1 gene are skin and brain. First, cholesterol sulfate plays an essential role in skin development and expression of the human SULT2B1b isoform, which is now recognized as a cholesterol sulfotransferase. Second, cholesterol sulfate levels are highest in skin. Third, the human fetal brain, as determined by RT-PCR, expresses only the SULT2B1a isoform. Finally, sulfated pregnenolone, which is produced most efficiently by the action of the SULT2B1a isoform, is now appreciated as an essential neurosteroid.
Mouse Ortholog of Human SULT2B1
Shimizu, Fuda, Yanei, Strott
We have cloned a mouse SULT2B1 cDNA (designated SULT2B1a) and found that it is homologous to a previously cloned mouse cDNA (designated SULT2B1b). The mouse SULT2B1 gene structure is identical to that of the human SULT2B1 gene, indicating that the gene is highly conserved. Thus, similar to the human SULT2B1 isoforms, the mouse counterparts arise from a single SULT2B1 gene by using an alternative exon 1 coupled with differential splicing. Therefore, as with the human SULT2B1 isoforms, the mouse SULT2B1 isoforms differ only at their amino termini. Furthermore, the mouse SULT2B1 isoforms, like human SULT2B1, have a predilection for cholesterol over DHEA, which is ordinarily sulfonated by the mouse SULT2A1 subfamily of hydroxysteroid sulfotransferases.
The differential expression of mouse SULT2B1a, SULT2B1b, and SULT2A1, along with their distinct substrate preferences, no doubt reflects their physiologic roles. Thus, the exclusive expression of SULT2B1a in the central nervous system and its preference for pregnenolone as a substrate are in keeping with the importance of pregnenolone sulfate as a neurosteroid; the exclusive expression of SULT2B1b in skin and its preference for cholesterol as a substrate are in keeping with the importance of cholesterol sulfate as a regulatory molecule in keratinocyte differentiation and development of the epidermal barrier; and the exclusive expression of SULT2A1 in the liver is in keeping with its role in general metabolism involving xenobiotics as well as endobiotics. Interestingly, the mouse SULT2B1 and SULT2A1 genes are differentially expressed during embryonic development, the former at all stages from E8.5 through E19 and the latter not until E19. The biologic and developmental significance of this observation remains to be elucidated. To this end, we aim to develop a gene knock-out mouse model, first disrupting the gene for SULT2B1 and then constructing double knock-out of the SULT2B1 and SULT2A1 genes.
PUBLICATIONS
- Fuda H, Lee YC, Shimizu C, Javitt N, Strott CA.
Mutational analysis of human hydroxysteroid sulfotransferase SULT2B1
isoforms reveals that exon IB of the SULT2B1 gene produces cholesterol
sulfotransferase, whereas exon IA yields pregnenolone sulfotransferase.
J Biol Chem. 2002;277:36161-36166.
- Fuda H, Shimizu C, Lee YC, Akita H, Strott CA. Characterization
and expression of human 3'-phosphoadenosine 5-phosphosulfate synthase
isoforms. Biochem J. 2002;365:497-504.
- Shimizu C, Fuda H, Lee YC, Strott CA. Transcriptional regulation
of human 3'-phosphoadenosine 5'-phosphosulfate synthase 2. Biochem
J. 2002;363:263-271.
- Strott CA. Dehydroepiandrosterone and cholesterol sulfotransferases.
In: Coughtrie MWH, Pacifici GM, eds. Human cytosolic sulfotransferases.
London;Taylor & Francis, 2002;in press.
- Strott CA. Sulfonation and molecular action. Endocr Rev. 2002;23:703-732.