Changing transcription start sites in H-type alpha(1,2)fucosyltransferase gene (FUT1) during differentiation of the human erythroid lineage

Emergence of an erythroid cell-specific regulatory region in ABO intron 1 attributable to A- or B-antigen expression on erythrocytes in Hominoidea

A- and B-antigens are present on red blood cells (RBCs) as well as other cells and secretions in Hominoidea including humans and apes such as chimpanzees and gibbons, whereas expression of these antigens on RBCs is subtle in monkeys such as Japanese macaques. Previous studies have indicated that H-antigen expression has not completely developed on RBCs in monkeys. Such antigen expression requires the presence of H-antigen and A- or B-transferase expression in cells of erythroid lineage, although whether or not ABO gene regulation is associated with the difference of A- or B-antigen expression between Hominoidea and monkeys has not been examined. Since it has been suggested that ABO expression on human erythrocytes is dependent upon an erythroid cell-specific regulatory region or the + 5.8-kb site in intron 1, we compared the sequences of ABO intron 1 among non-human primates, and demonstrated the presence of sites orthologous to the + 5.8-kb site in chimpanzees and gibbons, and their absence in Japanese macaques. In addition, luciferase assays revealed that the former orthologues enhanced promoter activity, whereas the corresponding site in the latter did not. These results suggested that the A- or B-antigens on RBCs might be ascribed to emergence of the + 5.8-kb site or the corresponding regions in ABO through genetic evolution.

Hypoxia inducible factor 1α down regulates cell surface expression of α1,2-fucosylated glycans in human pancreatic adenocarcinoma cells

The α1,2-fucosyltransferase activity in pancreatic tumors is much lower compared to normal pancreatic tissue. Here we show that hypoxia inducible factor (HIF) 1α is constitutively expressed in the pancreatic cancer cell lines Pa-Tu-8988S and Pa-Tu-8988T and suppresses the expression of the α1,2-fucosyltransferase genes FUT1 and FUT2. Down regulation of HIF-1α expression resulted in elevated FUT1 and FUT2 transcript levels and an increased expression of α1,2-fucosylated glycan structures on the surface of these cells. In conclusion, our data are the first to identify HIF-1α as a suppressor of FUT1/2 expression, thereby regulating α1,2-fucosylation of cell-surface glycans.

Transcriptional regulation of fucosyltransferase 1 gene expression in colon cancer cells

The α 1,2-fucosyltransferase I (FUT1) enzyme is important for the biosynthesis of H antigens, Lewis B, and Lewis Y. In this study, we clarified the transcriptional regulation of FUT1 in the DLD-1 colon cancer cell line, which has high expression of Lewis B and Lewis Y antigens, expresses the FUT1 gene, and shows α 1,2-fucosyltransferase (FUT) activity. 5'-rapid amplification of cDNA ends revealed a FUT1 transcriptional start site -10 nucleotides upstream of the site registered at NM_000148 in the DataBase of Human Transcription Start Sites (DBTSS). Using the dual luciferase assay, FUT1 gene expression was shown to be regulated at the region -91 to -81 nt to the transcriptional start site, which contains the Elk-1 binding site. Site-directed mutagenesis of this region revealed the Elk-1 binding site to be essential for FUT1 transcription. Furthermore, transfection of the dominant negative Elk-1 gene, and the chromatin immunoprecipitation (CHIp) assay, supported Elk-1-dependent transcriptional regulation of FUT1 gene expression in DLD-1 cells. These results suggest that a defined region in the 5'-flanking region of FUT1 is critical for FUT1 transcription and that constitutive gene expression of FUT1 is regulated by Elk-1 in DLD-1 cells.

ABO transcript levels in peripheral blood and erythropoietic culture show different allele-related patterns independent of the CBF/NF-Y enhancer motif and multiple novel allele-specific variations in the 5'- and 3'-noncoding regions

BACKGROUND

Mechanisms regulating the ABO gene are unclear, especially in the hematopoietic compartment. The number of 43-bp repeats in the CBF/NF-Y-binding enhancer region is considered to have a major influence on transcription.

STUDY DESIGN AND METHODS

Transcript levels in peripheral blood and in erythropoietic culture of CD34+ cells from marrow donors were measured with TaqMan assays. The 5'-regulatory region and 3'-downstream sequences were investigated to determine if allelic variations occur.

RESULTS

Surprisingly, transcripts from A(1) and A(2) alleles could not be detected in peripheral blood, although transcripts from B/O(1)/O(1v)/O(2) alleles were readily observed. Sequencing of approximately 4 kb upstream and 1.8 kb downstream of the coding region showed multiple novel allele-specific and allele-related motifs. No correlation between these sequence variations and transcript levels was found, however. Contradictory to the results with peripheral blood, in erythropoietic culture of CD34+ cells from healthy marrow donors transcripts from A(1) and A(2) alleles were found at higher levels than transcripts from B/O(1)/O(1v) alleles.

CONCLUSION

These data do not support previous suggestions that nonsense-mutated O(1)/O(1v) transcripts are eliminated first. Furthermore, our results contradict the notion that the number of repeats in the upstream CBF/NF-Y-binding enhancer region, which contains four 43-bp repeats in A(2)/B/O(1)/O(1v) but only one 43-bp unit in A(1)/O(2) alleles, determines the transcription rate. The reason for the remarkable discrepancy between blood and marrow remains to be elucidated.

Determinants of ABH expression on human blood platelets

Platelets express ABH antigens, which can adversely effect platelet transfusion recovery and survival in ABH-incompatible recipients. To date, there has been no large, comprehensive study comparing specific donor factors with ABH expression on platelet membranes and glycoconjugates. We studied ABH expression in 166 group A apheresis platelet donors by flow cytometry, Western blotting, and thin layer chromatography relative to donor age, sex, A1/A2 subgroup, and Lewis phenotype. Overall, A antigen on platelet membranes, glycoproteins, and glycosphingolipids was linked to an A1 red blood cell (RBC) phenotype. Among A1 donors, platelet ABH varied significantly between donors (0%-87%). Intradonor variability, however, was minimal, suggesting that platelet ABH expression is a stable, donor-specific characteristic, with 5% of A1 donors typing as either ABH high- or low-expressers. Group A2 donors, in contrast, possessed a Bombay-like phenotype, lacking both A and H antigens. Unlike RBCs, ABH expression on platelets may be determined primarily by H-glycosyltransferase (FUT1) activity. Identification of A2 and A1 low expressers may increase the availability and selection of crossmatched and HLA-matched platelets. Platelets from group A2 may also be a superior product for patients undergoing A/O major mismatch allogeneic progenitor cell transplantation.

Characterization of the promoter region of the human Galbeta1,3(4)GlcNAc alpha2,3-sialyltransferase III (hST3Gal III) gene

Multiple promoters are found in the hST3Gal IV, hST3Gal V and hST3Gal VI genes. These promoters may respond to different physiological signals and stimuli in different cell types. The multiple regulatory pathways of these ubiquitous sialyltransferases may need to be differentially modulated in various cell types. Here, we report transcriptional regulation of the hST3Gal III gene. 5'-RACE analysis determined that the transcription initiation sites map at -181 bp from the translation initiation site in all four cell lines (K-562, HT-29, PC-3 and HepG2) tested. Our results suggest that the hST3Gal III gene does not have multiple mRNAs, as have been identified for the hST3Gal IV, hST3Gal V and hST3Gal VI genes. The 5'-untranslated region was found to be divided into two exons, E1 and E2, indicating that the transcriptional regulation of hST3Gal III depends on the pIII promoter that exists 5'-upstream of exon E1. Luciferase assay results suggest that the nt -303 to -1 region is important for transcriptional activity of the hST3Gal III gene in all four cell lines tested. These results suggest that ubiquitous factors, such as Sp1, may be important for hST3Gal III gene expression.

Transcriptional regulation of human Galbeta1,3GalNAc/Galbeta1, 4GlcNAc alpha2,3-sialyltransferase (hST3Gal IV) gene in testis and ovary cell lines

The mRNA expression of sialyltransferase genes is regulated in a cell-type-specific manner. The mRNAs of human Galbeta1, 3GalNAc/Galbeta1, 4 GlcNAc alpha2,3-sialyltransferase gene (hST3Gal IV) consist of six isoforms, type A1, A2, B1, B2, B3, and BX. These mRNAs are transcribed from different promoters, pA, pB1, pB2, pB3, and pBX, respectively. Type B mRNAs are expressed in several cells, whereas type A mRNAs are specifically expressed in testis, ovary, and placenta, suggesting that pA promoter activity is especially high in these tissues. We show herein germ-cell specific transcriptional regulation of the hST3Gal IV pA promoter. Using a luciferase assay, pA promoter activity is shown to be high in testis and ovary cell lines. We identified the enhancer region of the pA promoter, located at nt -520 to -420. These results suggest that this element plays a critical role in germ-cell specific regulation of the pA promoter. The results of site-directed mutagenesis suggest that AP2 and c-Ets sites in this region are involved in pA promoter activity, which in turn suggests that the hST3Gal IV gene is regulated in a tissue-restricted fashion at the level of transcription.

Organization of the bovine alpha 2-fucosyltransferase gene cluster suggests that the Sec1 gene might have been shaped through a nonautonomous L1-retrotransposition event within the same locus

By referring to the split coding sequence of the highly conserved alpha 6-fucosyltransferase gene family (assumed to be representative of the common alpha 2 and alpha 6 fucosyltransferase gene ancestor), we have hypothesized that the monoexonic coding sequences of the present alpha 2-fucosyltransferase genes have been shaped in mammals by several events of retrotransposition and/or duplication. In order to test our hypothesis, we determined the structure of the three bovine alpha 2-fucosyltransferase genes (bfut1, bfut2, and sec1) and analyzed their characteristics compared with their human counterparts (FUT1, FUT2, and Sec1). We show that in mammals, a complex nonautonomous L1-retrotransposition event occurred within the locus of the alpha 2-fucosyltransferase ancestor gene itself. A consequence of this event was the processing in Catarrhini of a Sec1 pseudogene via several point mutations.

Gene structure and transcriptional regulation of human Gal beta1,4(3) GlcNAc alpha2,3-sialyltransferase VI (hST3Gal VI) gene in prostate cancer cell line

We describe transcriptional regulation of the human Gal beta1,4(3) GlcNAc alpha2,3-sialyltransferase VI (hST3Gal VI) gene. The 5'-RACE results indicated that two mRNA forms differ only in the 5'-untranslated region (types 1 and 2). The genomic structure shows that the transcriptional regulation of type 1 and type 2 mRNA depends on the P1 and P2 promoters, respectively. Northern blots of RNA derived from various human tissues showed that the expression level of type 2 mRNA is higher than that of type 1 in the prostate. To elucidate the molecular basis of hST3Gal VI gene expression, we isolated and functionally characterized the genomic region containing the P1 and P2 promoters of hST3Gal VI. The activity of the P2 promoter is much higher than that of the P1 promoter in the prostate adenocarcinoma cell line PC-3. The results suggested that the hST3Gal VI gene is expressed specifically by alternative promoter utilization and is regulated in a tissue-restricted fashion at the level of transcription.

The polymorphisms of fucosyltransferases

The alpha(1,2)fucosyltransferase Se enzyme regulates the expression of the ABH antigens in secretion. Secretors, who have ABH antigens in their saliva, have at least one functional Se allele in the FUT2 locus, while non-secretors, who fail to express ABH antigens in saliva, are homozygous for the non-functional se allele. Molecular analyses of the FUT2 polymorphism of various populations have indicated the ethnic specificity of null alleles: the null allele se(428) is a common Se enzyme-deficient allele in Africans and Caucasians but does not occur in Asians, whereas the null allele se(357,385) is specific to Asians. The gene frequency of se(428) or se(357,385) is about 0.5 in each respective population. Why the se(428) is absent in Asians is of interest. Also here, we describe the polymorphisms of the fucosyltransferase genes (FUT1, FUT3 and FUT6).

Identification of a novel FGF, FGF-21, preferentially expressed in the liver

We isolated cDNA encoding a novel FGF (210 amino acids) from mouse embryos. As this is the 21st documented FGF, we tentatively term it FGF-21. FGF-21 has a hydrophobic amino terminus ( approximately 30 amino acids), which is a typical signal sequence, and appears to be a secreted protein. The expression of FGF-21 mRNA in mouse adult tissues was examined by Northern blotting analysis. FGF-21 mRNA was most abundantly expressed in the liver, and also expressed in the thymus at lower levels. We also isolated human cDNA encoding FGF-21 (209 amino acids). Human FGF-21 is highly identical ( approximately 75% amino acid identity) to mouse FGF-21. Among human FGF family members, FGF-21 is most similar ( approximately 35% amino acid identity) to FGF-19.

Evolution of alpha 2-fucosyltransferase genes in primates: relation between an intronic Alu-Y element and red cell expression of ABH antigens

Coding sequences of the paralogous FUT1 (H), FUT2 (Se), and Sec1 alpha 2-fucosyltransferase genes were obtained from different primate species. Analysis of the primate FUT1-like and FUT2-like sequences revealed the absence of the known human inactivating mutations giving rise to the h null alleles of FUT1 and the se null alleles of FUT2. Therefore, most primate FUT1-like and FUT2-like genes potentially code for functional enzymes. The Sec1-like gene encodes for a potentially functional alpha 2-fucosyltransferase enzyme in nonprimate mammals, New World monkeys, and Old World monkeys, but it has been inactivated by a nonsense mutation at codon 325 in the ancestor of humans and African apes (gorillas, chimpanzees). Human and gorilla Sec1's have, in addition, two deletions and one insertion, respectively, 5' of the nonsense mutation leading to proteins shorter than chimpanzee Sec1. Phylogenetic analysis of the available H, Se, and Sec1 mammalian protein sequences demonstrates the existence of three clusters which correspond to the three genes. This suggests that the differentiation of the three genes is rather old and predates the great mammalian radiation. The phylogenetic analysis also suggests that Sec1 has a higher evolutionary rate than FUT2 and FUT1. Finally, we show that an Alu-Y element was inserted in intron 1 of the FUT1 ancestor of humans and apes (chimpanzees, gorillas, orangutans, and gibbons); this Alu-Y element has not been found in monkeys or nonprimate mammals, which lack ABH antigens on red cells. A potential mechanism leading to the red cell expression of the H enzyme in primates, related to the insertion of this Alu-Y sequence, is proposed.

Functional analysis of the 5'-flanking region of FTA for expression of rat GDP-L-fucose:beta-D-galactoside 2-alpha-L-fucosyltransferase

The tissue-specific and species-specific expression of the ABH antigens is well known among vertebrate species and it is regulated by the alpha(1,2)fucosyltransferase that forms the H antigen, a precursor of the A and B antigens. To investigate the mechanisms governing the tissue-specific and species-specific expression of this alpha(1,2)fucosyltransferase, we characterized the gene structure, including the promoter region, of FTA, a rat orthologous homolog of human FUT1 that encodes the H alpha(1, 2)fucosyltransferase and is responsible for the expression of the ABH antigens on human red blood cells. Northern blot and 5'-RACE analyses suggested that at least two forms of FTA mRNA (2.9 and 2.6 kb), which use alternative transcription start sites, are present in the cancer cell lines RCN-9 (rat colon cancer) and PC12 (rat pheochromocytoma), whereas only the 2.6 kb form was detected in normal colon, stomach and pancreas. Transcriptional activity of the 5'-flanking sequence, which contains three putative Sp1-binding sites, but lacks both TATA and CAAT boxes, was examined. Transient transfection experiments of promoter-reporter gene constructs showed high promoter activity in RCN-9, PC12 and human colon cancer (WiDr) cell lines, weak activity in human vascular endothelial (ECV304) cells and no activity in human erythroleukemia (HEL) cells. The results suggest that the 5'-flanking region of FTA contains a tissue-specific promoter. Deletional analysis of the 5'-flanking sequence revealed regions containing cell-type-specific positive acting element(s) and negative regulatory element(s), which are related to the promoter activity.

A novel zymogen granule protein (ZG29p) and the nuclear protein MTA1p are differentially expressed by alternative transcription initiation in pancreatic acinar cells of the rat

Using a polyclonal antibody against purified zymogen granule membrane components from rat pancreas a cDNA coding for the 29 kDa protein (ZG29p) was identified by immunoscreening of a hormonally stimulated pancreas cDNA library. Western blot analysis suggests that ZG29p is a pancreas-specific protein and immunofluorescence shows that ZG29p is mainly associated with zymogen granules. Analysis of subcellular fraction applying immunoblotting revealed that ZG29p was localized mainly in the soluble fraction of zymogen granules and in a Golgi- and RER-enriched fraction, but was absent from the cytosol. In isolated zymogen granule content ZG29p was associated with protein complexes containing amylase as main constituent. The cDNA coding for ZG29p is homologous to the C-terminal region of the candidate metastasis-associated gene mta1. Northern blot analysis and RT-PCR showed that no MTA1 mRNA is present in pancreas from fasted rats and in the rat pancreas carcinoma cell line AR4-2J in its protodifferentiated state. Although no ZG29p specific mRNA was seen in the northern blot analysis, RT-PCR showed that ZG29p was expressed under both non-stimulated and stimulated conditions. The expression of MTA1 was up-regulated in the pancreas by endogenous cholecystokinin release and in AR4-2J after induction of cellular differentiation by dexamethasone. Western blotting and immunofluorescense studies indicated that MTA1p is localized in the nucleus in all tissues studied. Using genomic DNA in PCR analysis it was shown that two short introns are present flanking the sequences of the 5'end of ZG29p cDNA. One intron contains consensus elements required for pancreas specific transcription initiation, suggesting that MTA1 and ZG29 are differentially expressed by alternative transcription initiation in the pancreas. The localisation of MTA1p in the nucleus of most cell types could signify a general role in gene regulation, while the cell type specific and exclusive expression of ZG29p in pancreatic acinar cells could indicate a role in granule formation.