Construction of a human glycogene library and comprehensive functional analysis

Eighteen years have passed after the first mammalian glycosyltransferase was cloned. At the beginning of April, 2001, 110 genes for human glycosyltransferases, including modifying enzymes for carbohydrate chains such as sulfotransferases, had been cloned and analyzed. We started the Glycogene Project (GG project) in April 2001, a comprehensive study on human glycogenes with the aid of bioinformatic technology. The term glycogene includes the genes for glycosyltransferases, sulfotransferases adding sulfate to carbohydrates and sugar-nucleotide transporters, etc. Firstly, as many novel genes, which are the candidates for glycogenes, as possible were searched using bioinformatic technology in databases. They were then cloned and expressed in various expression systems to detect the activity for carbohydrate synthesis. Their substrate specificity was determined using various acceptors.     

The Drosophila melanogaster brainiac protein is a glycolipid-specific beta 1,3N-acetylglucosaminyltransferase

Mutations at the Drosophila melanogaster brainiac locus lead to defective formation of the follicular epithelium during oogenesis and to neural hyperplasia. The brainiac gene encodes a type II transmembrane protein structurally similar to mammalian beta1,3-glycosyltransferases. We have cloned the brainiac gene from D. melanogaster genomic DNA and expressed it as a FLAG-tagged recombinant protein in Sf9 insect cells. Glycosyltransferase assays showed that brainiac is capable of transferring N-acetylglucosamine (GlcNAc) to beta-linked mannose (Man), with a marked preference for the disaccharide Man(beta1,4)Glc, the core of arthro-series glycolipids. The activity of brainiac toward arthro-series glycolipids was confirmed by showing that the enzyme efficiently utilized glycolipids from insects as acceptors whereas it did not with glycolipids from mammalian cells. Methylation analysis of the brainiac reaction product revealed a beta1,3 linkage between GlcNAc and Man, proving that brainiac is a beta1,3GlcNAc-transferase. Human beta1,3GlcNAc-transferases structurally related to brainiac were unable to transfer GlcNAc to Man(beta1,4)Glc-based acceptor substrates and failed to rescue a homozygous lethal brainiac allele, indicating that these proteins are paralogous and not orthologous to brainiac.     

Glycosylation diseases: Quo vadis?

About 250 to 500 glycogenes (genes that are directly involved in glycan assembly) are in the human genome representing about 1–2% of the total genome. Over 40 human congenital diseases associated with glycogene mutations have been described to date. It is almost certain that the causative glycogene mutations for many more congenital diseases remain to be discovered. Some glycogenes are involved in the synthesis of only a specific protein and/or a specific class of glycan whereas others play a role in the biosynthesis of more than one glycan class. Mutations in the latter type of glycogene result in complex clinical phenotypes that present difficult diagnostic problems to the clinician. In order to understand in biochemical terms the clinical signs and symptoms of a patient with a glycogene mutation, one must understand how the glycogene works. That requires, first of all, determination of the target protein or proteins of the glycogene followed by an understanding of the role, if any, of the glycogene-dependent glycan in the functions of the protein. Many glycogenes act on thousands of glycoproteins. There are unfortunately no general methods to identify all the potentially large number of glycogene target proteins and which of these proteins are responsible for the mutant phenotypes. Whereas biochemical methods have been highly successful in the discovery of glycogenes responsible for many congenital diseases, it has more recently been necessary to use other methods such as homozygosity mapping. Accurate diagnosis of many recently discovered diseases has become difficult and new diagnostic procedures must be developed. Last but not least is the lack of effective treatment for most of these children and of animal models that can be used to test new therapies.     

Induction of differentiation of undifferentiated cells into pancreatic beta cells in vertebrates

The beta cells of the pancreatic islets, which maintain glucose homeostasis by secreting insulin, are important cells for sustaining life. In recent years, islet transplantation has been performed as a treatment for type I diabetes. Since there are not enough donors for patients awaiting transplantation, beta cells grown in vitro are expected to be utilized as a substitute for islets. To obtain the cells with properties of human beta cells, it is necessary to understand the process by which human pancreatic islets are formed, as well as their structural characteristics. By using undifferentiated cells, such as Xenopus laevis animal caps and mouse ES cells, pancreatic tissue has shown to be able to be induced in vitro. Various attempts have been made to obtain human beta cells from human ES/iPS cells. Versatile methods have been developed and improved efficiency has been achieved by the use of low molecular weight compounds, but the challenge remains to prevent tumor formation and achieve functional maturation. Inducing the differentiation of somatic stem cells into insulin-producing cells has also brought us closer to clinical application. There are still many challenges related to the practical use of beta cells derived from undifferentiated cells, such as the development of methods to substitute these cells for host beta cells, standardization of the treatment protocol, quality control, and confirmation of safety. Research on the methods of inducing undifferentiated cells to differentiate into beta cells has shown definite progress, suggesting that cell therapy for diabetes may become a preferred therapeutic option over islet transplantation.     

Chromosomal localization of the genes encoding two forms of the G protein beta polypeptide, beta 1 and beta 3, in man

The signal-transducing G proteins are heterotrimers composed of three subunits, alpha, beta, and gamma. Multiple distinctive forms of the alpha, beta, and gamma subunits, each encoded by a distinct gene, have been described. To investigate further the structural diversity of the beta subunits, we recently cloned and characterized a novel cDNA encoding a third form of the G protein beta subunit, which we have termed beta 3. The protein corresponding to beta 3 has not yet been identified. The three forms of the beta subunit show 81-90% amino acid sequence identity. Previous studies had localized the human genes for the beta 1 and beta 2 subunits to chromosomes 1 and 7, respectively. The present studies were designed to determine whether the gene encoding beta 3 is linked to either the beta 1 or the beta 2 gene. Genomic DNA was isolated from a panel of rodent-human hybrid cell lines and analyzed by hybridization to cDNAs for beta 1 and beta 3. Discordancy analysis allowed assignment of the beta 3 gene to chromosome 12 and confirmed the previous assignment of the beta 1 gene to chromosome 1. These results were confirmed and extended by using in situ chromosome hybridization, which permitted the regional localization of the beta 1 gene to 1pter----p31.2 and the beta 3 gene to 12pter----p12.3. Digestion of human genomic DNA with 10 restriction enzymes failed to disclose a restriction fragment length polymorphism for the beta 3 gene. These data indicate that there is considerable diversity in the genomic organization of the beta subunit family.     

Glycomic alterations are associated with multidrug resistance in human leukemia

Correlations of disease phenotypes with glycosylation changes have been analyzed intensively in tumor biology field. In this study we describe glycomic alterations of multidrug resistance in human leukemia cell lines. Using multiple glycan profiling tools: real-time PCR for quantification of glycogenes, FITC-lectin binding for glycan profiling, and mass spectrometry for glycan composition, we compared the glycomics of drug-resistant K562/ADR cells with parental K562 line. The results showed that the expression of glycogenes, glycan profiling and N-glycan composition were different in K562/ADR cells, as compared with those in K562 cells, whereas O-glycans of the two cell lines showed no different mass spectra. Further analysis of the N-glycan regulation by way of tunicamycin application or PNGase F treatment in K562/ADR cells showed partial inhibition of biosynthesis and increased sensitivity to chemotherapeutic drugs in vitro. We targeted glycogene B3GNT8 and ST8SIA4, which were over-expressed in K562/ADR cells, and silenced the expression levels of two glycogenes after using RNA interference approach. The results showed that the silencing of B3GNT8 or ST8SIA4 in K562/ADR cells resulted in increased chemosensitivity to anti-tumor drugs. In conclusion, glycomic alterations are responsible for the overcoming multidrug resistance in human leukemia therapy and the N-linked oligosaccharides are associated with the drug resistance of cancer cells.     

Bioinformatics for comprehensive finding and analysis of glycosyltransferases

Bioinformatics is a very powerful tool in the field of glycoproteomics as well as genomics and proteomics. As a part of the Glycogene Project (GG project), we have developed a novel bioinformatics system for the comprehensive identification and in silico cloning of human glycogenes. Using our system, a total of 105 candidate human glycogenes were identified and then engineered for heterologous expression. Of these candidates, 38 recombinant proteins were successfully identified for their enzyme activity and substrate specificity. We also classified 47 out of 60 carbohydrate-active enzyme glycosyltransferase families into 4 superfamilies using the profile Hidden Markov Model method. On the basis of our classification and the relationship between glycosylation pathways and superfamilies, we propose the evolution of glycosyltransferases.     

Mapping of genes for inhibin subunits alpha, beta A, and beta B on human and mouse chromosomes and studies of jsd mice

Inhibin (INH) is a gonadal glycoprotein hormone that regulates pituitary FSH secretion and may also play a role in the regulation of androgen biosynthesis. There are two forms of inhibin that strongly inhibit pituitary FSH secretion. These share the same alpha subunit that is covalently linked to one of two distinct beta subunits (beta A or beta B). However, dimers of two beta subunits are potent stimulators of FSH synthesis and release in vitro. The beta subunits share extensive sequence similarity with transforming growth factor beta. Recently isolated cDNAs for all three inhibin subunits have been used to map their cognate loci on human and mouse chromosomes by Southern blot analysis of somatic cell hybrid DNAs and by in situ hybridization. INH alpha and INH beta B genes were assigned to human chromosome 2, regions q33----qter and cen----q13, respectively, and to mouse chromosome 1. The INH beta A locus was mapped to human chromosome 7p15----p14 and mouse chromosome 13. The region of mouse chromosome 1 that carries other genes known to have homologs on human chromosome 2q includes the jsd locus (for juvenile spermatogonial depletion). Adult jsd/jsd mice have elevated levels of serum FSH and their testes are devoid of spermatogonial cells. The possibility that the mutation in jsd involves the INH alpha or INH beta B gene was investigated by Southern blotting of DNA from jsd/jsd mice, and no major deletions or rearrangements were detected.     

Differences in the activation mechanism between the alpha and beta subunits of human meprin

Meprins are zinc-endopeptidases of the astacin family, which are expressed as membrane-bound or secreted forms in renal and intestinal brush-border membranes of mouse, rat and man. There are two types of meprin subunits, alpha and beta, which form disulfide-bonded homo- and heterodimers; further oligomerization is mediated by non-covalent interactions. Both subunits are translated as proenzymes that have to be activated by removal of an N-terminal propeptide. In the gut, the most probable activator is trypsin. In addition, plasmin has been shown to activate the human alpha subunit in colorectal cancer tissue. In the present study we have overexpressed the human meprin alpha subunit and a His-tagged soluble tail-switch-mutant of meprin beta in Baculovirus-infected insect cells. The recombinant homo-oligomeric proteins were purified by gel filtration and affinity chromatography with yields of up to 10 mg/l cell culture medium and analyzed with regard to their activation mechanism. While both alpha and beta homo-oligomers are activated by trypsin, only meprin alpha homo-oligomers are processed to their mature form by plasmin. These results indicate a different accessibility of the propeptide in meprin homo-oligomers and suggest an explanation for the appearance of meprin hetero-oligomers consisting of active alpha, but latent beta subunits.     

Control of beta globin genes

The developmental changes in expression of the beta like genes from embryonic to adult stages of human life are controlled at least partially at the level of the promoter sequences of these genes and their binding factors, and competition for promoter specific interactions with the locus control region (LCR). In recent years, the control of beta globin genes has also been investigated at the level of chromatin structure involving the chemical modification of histones and their remodelling by DNA dependent ATPases (SMARCA) containing protein complexes. The role of intergenic RNA is also being investigated with renewed interest. Although a wealth of information on the structure/function relationship of the LCR and globin promoters has been gathered over more than two decades, the fundamental nature of the control of these genes at the molecular level is still not completely understood. In the following pages, we intend to briefly describe the progress made in the field and discuss future directions.     

Growth factors and beta cell replication

Recent studies have demonstrated that human islet allograft transplantation can be a successful therapeutic option in the treatment of patients with Type I diabetes. However, this impressive recent advance is accompanied by a very important constraint. There is a critical paucity of pancreatic islets or pancreatic beta cells for islet transplantation to become a large-scale therapeutic option in patients with diabetes. This has prompted many laboratories around the world to invigorate their efforts in finding ways for increasing the availability of beta cells or beta cell surrogates that potentially could be transplanted into patients with diabetes. The number of studies analyzing the mechanisms that govern beta cell proliferation and growth in physiological and pathological conditions has increased exponentially during the last decade. These studies exploring the role of growth factors, intracellular signaling molecules and cell cycle regulators constitute the substrate for future strategies aimed at expanding human beta cells in vitro and/or in vivo after transplantation. In this review, we describe the current knowledge on the effects of several beta cell growth factors that have been shown to increase beta cell proliferation and expand beta cell mass in vitro and/or in vivo and that they could be potentially deployed in an effort to increase the number of patients transplanted with islets. Furthermore, we also analyze in this review recent studies deciphering the relevance of these specific islet growth factors as physiological and pathophysiological regulators of beta cell proliferation and islet growth.     

Preferential expression of reg I beta gene in human adult pancreas

In human pancreas two genes, reg I alpha and reg I beta, have been characterized but only the reg I alpha protein has been isolated from human pancreatic secretion. To examine their respective physiological roles in fetal and adult pancreas we have compared the patterns of gene expression using a specific RT-PCR method. No progressive evolution in the two mRNAs levels was observed during fetal development (16--41 weeks). A discoordinate expression of the two genes was found with a higher level of reg I alpha mRNA in fetus and a higher level of regI beta in adult. In addition, if reg I alpha mRNA level was correlated with the expression of genes encoding exocrine proteins in adults, reg I beta mRNA level presented no correlation with any ductular, endocrine, or exocrine gene expression. In human pancreatic cell lines we showed the only expression of reg I beta gene and protein. All these data suggest that the two reg genes and proteins could play different roles in the pancreas.     

Construction of a genome library from a beta-0-thalassemic individual from Ferrara: characterization of clones containing beta globin genes

A library of genomic DNA was prepared from a patient with beta o Ferrara thalassaemia: random human DNA fragments (15 - 20 Kb) have been joined to phage lambda vectors and cloned has viable phage particles (4). 4x10(5) phages have been screened for their content in beta globin gene sequences, using a human beta cDNA plasmid (5) as hybridization probe. Five positive clones have been isolated and characterized by restriction endonuclease cleavage analysis and by the hybridization experiments. The results obtained allow the precise localization of the human fragments inside the beta like globin gene cluster (6). The comparison of the thalassaemic fragments with the normal DNA (6 - 7) shows two different restriction endonuclease sites, for Xba I and Eco RI respectively, downstream from the human beta globin gene.     

Human oestrogen receptors: differential expression of ER alpha and beta and the identification of ER beta variants

Two structurally related subtypes of oestrogen receptor (ER), known as alpha (ER alpha, NR3A1) and beta (ER beta, NR3A2) have been identified. ER beta mRNA and protein have been detected in a wide range of tissues including the vasculature, bone, and gonads in both males and females, as well as in cancers of the breast and prostate. In many tissues the pattern of expression of ER beta is distinct from that of ER alpha. A number of variant isoforms of the wild type beta receptor (ER beta 1), have been identified. In the human these include: (1). use of alternative start sites within the mRNA leading to translation of either a long (530 amino acids, hER beta 1L) or a truncated form (487aa hER beta 1s); (2). deletion of exons by alternative splicing; (3). formation of several isoforms (ER beta 2-beta 5) due to alternative splicing of exons encoding the carboxy terminus (F domain). We have raised monoclonal antibodies specific for hER beta1 as well as to three of the C terminal isoforms (beta2, beta 4 and beta 5). Using these antibodies we have found that ER beta 2, beta 4 and beta 5 proteins are expressed in nuclei of human tissues including the ovary, placenta, testis and vas deferens.In conclusion, in addition to the differential expression of full length ER alpha and ER beta a number of ER variant isoforms have been identified. The impact of the expression of these isoforms on cell responsiveness to oestrogens may add additional complexity to the ways in which oestrogenic ligands influence cell function.     

Generation of beta cells from human pluripotent stem cells: Potential for regenerative medicine

The loss of beta cells in Type I diabetes ultimately leads to insulin dependence and major complications that are difficult to manage by insulin injections. Given the complications associated with long-term administration of insulin, cell-replacement therapy is now under consideration as an alternative treatment that may someday provide a cure for this disease. Over the past 10 years, islet transplantation trials have demonstrated that it is possible to replenish beta cell function in Type I diabetes patients and, at least temporarily, eliminate their dependency on insulin. While not yet optimal, the success of these trials has provided proof-of-principle that cell replacement therapy is a viable option for treating diabetes. Limited access to donor islets has launched a search for alternative source of beta cells for cell therapy purposes and focused the efforts of many investigators on the challenge of deriving such cells from human embryonic (hESCs) and induced pluripotent stem cells (hiPSCs). Over the past five years, significant advances have been made in understanding the signaling pathways that control lineage development from human pluripotent stem cells (hPSCs) and as a consequence, it is now possible to routinely generate insulin producing cells from both hESCs and hiPSCs. While these achievements are impressive, significant challenges do still exist, as the majority of insulin producing cells generated under these conditions are polyhormonal and non functional, likely reflecting the emergence of the polyhormonal population that is known to arise in the early embryo during the phase of pancreatic development known as the 'first transition'. Functional beta cells, which arise during the second phase or transition of pancreatic development have been generated from hESCs, however they are detected only following transplantation of progenitor stage cells into immunocompromised mice. With this success, our challenge now is to define the pathways that control the development and maturation of this second transition population from hPSCs, and establish conditions for the generation of functional beta cells in vitro.     

Cytokine regulation of IL-1 beta gene expression in the human polymorphonuclear leukocyte

Although recently polymorphonuclear leukocytes (PMN) have been identified as producers of IL-1 beta in response to LPS and granulocyte/monocyte colony stimulating factor, little is known regarding the ability of other cytokines to induce the production of IL-1 beta in the PMN. Inasmuch as IL-1 and TNF have been shown to be important priming agents, as well as agents that induce migration of PMN, we investigated their effect on IL-1 beta gene expression in human peripheral blood PMN. In the present study, we demonstrate that human peripheral blood PMN produce IL-1 beta in response to IL-1 alpha, IL-1 beta, and TNF-alpha. Control (unstimulated) human PMN had virtually undetectable levels of IL-1 beta mRNA. Either IL-1 beta or TNF, induced PMN to transiently express IL-1 beta mRNA with peak expression at 1 h, returning to untreated levels by 2 h. A dose response indicated that as little as 0.05 ng/ml of IL-1 beta or TNF resulted in IL-1 beta induction, with maximal effects at 1 ng/ml of IL-1 beta and 5 ng/ml of TNF. IL-1 alpha or IL-1 beta exhibited similar dose responses in IL-1 beta mRNA induction. Inasmuch as cytokines have been shown to have synergistic effects in cell function studies, we induced PMN with a combination of maximally effective doses of TNF plus IL-1 beta. They demonstrated a cooperative effect on IL-1 beta gene expression, in that mRNA levels were sustained for three hours. IL-1 beta Ag expression, as measured by ELISA, paralleled IL-1 beta mRNA expression with cell associated peak levels at 2 to 4 h. IL-1 beta Ag levels in PMN lysates and supernatants correlated with IL-1 beta mRNA levels, i.e., TNF + IL-1 greater than TNF greater than IL-1. Thus, these studies represent the first demonstration of IL-1 and TNF induction of IL-1 beta gene expression in the PMN. Furthermore, the time course of induction is unique to the PMN, with peak induction of mRNA at 1 h, which is consistent with the short lived nature of these cells in inflammatory lesions.