Glycobiology Advance Access originally published online on October 19, 2005
Glycobiology 2006 16(2):117-131; doi:10.1093/glycob/cwj048
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
A focused microarray approach to functional glycomics: transcriptional regulation of the glycome
4 Department of Molecular Biology and Molecular and Experimental Medicine, The Scripps Research Institute, La Jolla, CA 92037; 5 DNA Microarray Core Facility, The Scripps Research Institute, La Jolla, CA 92037; 6 Imperial College London, South Kensington Campus, London SW7 2AZ, UK; 7 AIST, Tsukuba 305–8568, Japan; and 8 University of California San Diego, La Jolla, CA 92037
1 These authors contributed equally to this work.
2 To whom correspondence should be addressed; e-mail: jpaulson{at}scripps.edu
3 Present address: Nestlé Research Centre, Department of Nutrition and Health, Vers chez les Blanc, 1000 Lausanne 26, Switzerland
Received on July 15, 2005; revised on September 15, 2005; accepted on October 11, 2005
 |
Abstract |
|---|
 Top Abstract IntroductionResultsDiscussionMaterials and methodsAcknowledgmentsReferences |
|---|
Glycosylation is the most common posttranslational modification of proteins, yet genes relevant to the synthesis of glycan structures and function are incompletely represented and poorly annotated on the commercially available arrays. To fill the need for expression analysis of such genes, we employed the Affymetrix technology to develop a focused and highly annotated glycogene-chip representing human and murine glycogenes, including glycosyltransferases, nucleotide sugar transporters, glycosidases, proteoglycans, and glycan-binding proteins. In this report, the array has been used to generate glycogene-expression profiles of nine murine tissues. Global analysis with a hierarchical clustering algorithm reveals that expression profiles in immune tissues (thymus [THY], spleen [SPL], lymph node, and bone marrow [BM]) are more closely related, relative to those of nonimmune tissues (kidney [KID], liver [LIV], brain [BRN], and testes [TES]). Of the biosynthetic enzymes, those responsible for synthesis of the core regions of N- and O-linked oligosaccharides are ubiquitously expressed, whereas glycosyltransferases that elaborate terminal structures are expressed in a highly tissue-specific manner, accounting for tissue and ultimately cell-type-specific glycosylation. Comparison of gene expression profiles with matrix-assisted laser desorption ionization-time of flight (MALDI-TOF) profiling of N-linked oligosaccharides suggested that the
1-3 fucosyltransferase 9, Fut9, is the enzyme responsible for terminal fucosylation in KID and BRN, a finding validated by analysis of Fut9 knockout mice. Two families of glycan-binding proteins, C-type lectins and Siglecs, are predominately expressed in the immune tissues, consistent with their emerging functions in both innate and acquired immunity. The glycogene chip reported in this study is available to the scientific community through the Consortium for Functional Glycomics (CFG) (http://www.functionalglycomics.org). Key words: Fut9 / glycomics / glycosyltransferase / lectin / microarray
|
Introduction |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsReferences |
|---|
Glycans occur as modifications of proteins (glycoproteins and proteoglycans) or lipids (glycosphingolipids) and form the core structure of the glycophospholipid (GPI) anchors by which many proteins are anchored to the membrane (Spiro, 2002
; Taylor and Drickamer, 2003). Glycosylation is the most common form of posttranslational modification of proteins, with over half of all proteins estimated to contain one or more glycan chains (Apweiler et al., 1999; Wells and Hart, 2003). The roles of glycans are diverse. They contribute to the folding and conformational stability of many proteins (Wang et al., 1996; Bosques et al., 2004; Helenius and Aebi, 2004), mediate host–pathogen interactions and aspects of innate immunity (Karlsson, 1995; Underhill, 2003; Colmenares et al., 2004; Smith and Helenius, 2004), and serve as ligands for glycan-binding proteins that mediate cell trafficking, cell adhesion, and cell signaling (Crocker, 2002; Esko and Selleck, 2002; Lowe, 2002; Rabinovich et al., 2002; Wells and Hart, 2003).
The glycome by definition comprises all glycan structures produced by an organism, whereas glycomics encompasses the study of glycan structure, biosynthesis, and function (von der Lieth et al., 2004). Although glycomics emerged as a separate discipline from proteomics, the two will become increasingly intertwined through the inextricable link between structure and function of their respective "omes." In contrast to the genomics and proteomics fields, in which the database and molecular tools are highly advanced, comparable tools for glycomics are in the formative stages (Henrissat and Bairoch, 1996; Cooper et al., 2003; Kikuchi et al., 2003; Taylor and Drickamer, 2003; Kanehisa et al., 2004; Lutteke et al., 2004). Current glycan databases have assembled collections of thousands of unique structures, and the entire glycome, while certainly finite, will presumably number in the tens of thousands. In principle, because all the information necessary for synthesis of the glycome is contained within the genome, the current genome and proteome databases are relevant to glycan structure. However, glycan structure is determined by the concerted action of numerous genes that code for glycosyltransferases, glycosidases, and other enzymes that synthesize and remodel glycan chains, as well as accessory enzymes involved in the synthesis and transport of nucleotide sugars. In addition, some glycosyltransferases are known to be differentially expressed in various cells, contributing to cell-type-specific glycosylation. Thus, prediction of glycan structures produced by a cell not only requires a knowledge of which of these enzymes is expressed, but also how they work together in a biosynthetic network to assemble glycans of defined structures resulting from the specificities of the enzymes themselves.
As a step toward obtaining global information relating to glycan biosynthesis, structure, and function, we have constructed a focused gene microarray (glycogene-chip v1) using the Affymetrix technology. The chip contains families of 436 human and 285 murine genes, termed here "glycogenes," coding for proteins responsible for glycan synthesis and glycan recognition, including glycosyltransferases and other glycan processing enzymes, enzymes relating to nucleotide synthesis and transport, proteoglycans, and glycan-binding proteins. Unlike large-scale commercial chips, it includes genes not currently represented and has been highly annotated by experts of each gene family, to facilitate interrogation of results with respect to gene function.
In this report, the chip has been used to evaluate the expression of glycogenes in nine murine tissues to establish the degree to which the genes are detected and determine whether it can serve as a useful tool for analysis of glycan synthesis and function. Remarkably, despite the fact that glycosyltransferases are expressed at only to 0.01–0.00001% of the cellular protein (Weinstein et al., 1982; Ju et al., 2002), over 90% of all glycogenes were detected in at least one tissue. Enzymes involved in the synthesis of core regions of glycan chains were constitutively expressed, whereas those involving terminal regions were differentially expressed, providing global support for the concept that expression of these genes is a major factor in determining cell-type-specific glycosylation (Fukui et al., 2002; Sutton-Smith et al., 2002; Wang et al., 2002). Moreover, some features of N-linked glycan profiles obtained by matrix-assisted laser desorption ionization-time of flight mass spectrometry (MALDI-TOF MS) were reflected in the expression patterns of several glycosyltransferase genes. The results suggest that this highly annotated microarray will provide a useful tool for the analysis of the regulated synthesis of glycans and glycan function, and several reports using the glycogene-chip array for analysis of limited subsets of genes have appeared (Brown et al., 2003; Smith et al., 2005).
|
Results |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsReferences |
|---|
Glycogene expression profiles segregate tissues according to their biological function
Following optimization and validation of expression analysis using the glycogene-chip v1 (see Materials and methods), we first surveyed nine murine tissues to evaluate the degree to which glycogenes are detected. To take into account biological variability as well as variability contributed by tissue-collection procedures, array analysis was performed on RNA samples extracted from three independently prepared tissue samples. Each tissue sample was pooled from three age- and sex-matched littermates. The hybridization .CEL files and dChip expression values for each of the experiments performed are available at the Consortium for Functional Glycomics (CFG) Web site and at gene expression omnibus (GEO) (http://www.ncbi.nlm.nih.gov/geo/) under accession number GSE2781 [NCBI GEO] . Ninety percent of all glycogenes were detected as present in at least one tissue. There are several possible reasons why a probe set would fail to indicate the presence of a transcript in any of the tissues surveyed, including lack or low level of transcript expression, and faulty probe set design caused by incomplete sequence information or ineffective probe selection.
To estimate the degree of relationship among the nine tissues of glycogene expression, we extracted the probe sets falling in the glycosyltransferase, glycan-binding protein, and proteoglycan categories and performed an unsupervised hierarchical cluster analysis for each of these groups of genes independently, using the glycogene array data for every replicate (Figure 1). The high correlation (short branches) within each of the three replicates for each tissue shows that there was little variability among samples. In the three cases, the clustering separates the immune tissues (thymus [THY], spleen [SPL], lymph node, and bone marrow [BM]) from the nonimmune tissues (kidney [KID], liver [LIV], brain [BRN], and testes [TES]), with lung (LNG) clustering with the immune tissues in the case of proteoglycans. In particular, secondary immune organs (lymph nodes and SPL) are more closely related between them than to primary organs (BM and THY). We conclude that organs with similar functions have similar expression of glycogenes, both from capability of synthesizing specific glycans and from lectin expression, which implies the association of specific glycosylation patterns to their biological activities.
|
Expression patterns of glycan transferases involved in the synthesis of glycoprotein and glycolipid glycans
Core sequences
Each class of glycoprotein and glycolipid glycans is characterized by common core structures, which are variously elongated and terminated, comprising the basis for structural diversity of glycans. Figure 2 illustrates representative structures of N- and O-linked glycans of glycoproteins, and glycolipids, with the common core element indicated in the dashed box. For analysis of glycosyltransferase expression, we have grouped the glycosyltransferases represented on the chip with respect to their role in the synthesis of the core structure common to glycans in all cells or terminal glycosylation sequences which vary from cell type to cell type (Tables I and II). For this purpose, we have excluded other families of genes involved in glycan synthesis including glycosyltransferases involved in chain elongation (galactosyl- and N-acetylglucosaminyl transferases), and enzymes involved in the synthesis and transport of nucleotide sugars. A complete listing of the raw data can be found at http://www.functionalglycomics.org.
|
|
|
As summarized in Table I, enzymes involved in the synthesis of the core sequences of N-glycans are expressed at similar levels in every tissue and can therefore be considered as "housekeeping" genes, in keeping with the fact that N-linked glycosylation is found in all eukaryotic cells.
The only sugar common to all major glycans of the "mucin type" O-linked glycans is the GalNAcThr/Ser linkage, which is synthesized by a family of polypeptide N-acetylgalactosaminyl transferases (ppGalNAcT) that has grown to nearly 20 members in recent years (Ten Hagen et al., 2003; Young et al., 2003). Because the specificities of these enzymes in general have broad overlaps, one or more members of this family are sufficient to allow O-linked glycosylation. For the seven members of the family represented on this array, each tissue expresses 2–5 members. This is consistent with the ubiquitous expression of O-linked glycans given the overlapping specificity of these enzymes for initiation of O-linked glycosylation on the polypeptide chain (Bennett et al., 1996; Zara et al., 1996; Wandall et al., 1997).
For glycolipids, two classes have been described that have in common the first sugar added to the ceramide (lipid) moiety. The glucosyl ceramides are ubiquitously expressed in all cells with glycan chains of various size and diverse sequence. In contrast, galactosyl ceramide, or cerebroside, is not extended except for addition of sulfate and exhibits restricted expression to a few tissues (Varki, 1999). In keeping with the known distribution of these glycolipids, the glucosyltransferase that initiates the synthesis of the glucosylceramides is ubiquitously expressed, whereas the galactosyltransferase that forms the galactosylceramides is restricted to BRN, KID, and TES (Table I).
Terminal sequences
Biosynthetic elaboration of the cores structures to complete the terminal sequences gives rise to structural variation that differs from cell to cell. This is illustrated in Figure 3 for N-linked glycans, where the MALDI-MS profiles of the total extracted N-linked glycans of various tissues give a distinct spectrum of glycan structures. Each tissue exhibits terminal sequences that are a result of differential expression of the glycosyltransferases expressed in that cell. Not surprisingly, therefore, glycan transferases involved in terminal glycosylation exhibit highly variable expression from tissue to tissue, as illustrated in Table II for sialyltransferases (STs), fucosyltransferases, and sulfotransferases. No two tissues exhibit expression of the same complement of enzymes, and conversely the variable expression of each member of the family appears to be independent of the others.
|
The STs constitute a family of ~20 members (Harduin-Lepers et al., 2001). Together they attach sialic acids at terminal positions of glycoprotein and glycolipid glycans. Their names reflect the glycan linkage they form, NeuAc2–3Gal (ST3Gal), NeuAc2–6Gal (ST6Gal), NeuAc2–6GalNAc (ST6GalNAc), and NeuAc2–8NeuAc (STSia8). Of those represented on the chip, seven are known to attach sialic acid on N-linked glycans (ST3Gal III, IV, and VI; ST6Gal I; and ST8Sial II, III, and IV). As seen in the MALDI-TOF glycan profiles, sialylation of N-linked glycans is a predominant feature of all tissues examined, except KID (Figure 3). The NeuAc2–3/6Gal sequence that terminates these structures is caused by the expression of the ST3Gal and ST6Gal enzymes. No NeuAc–NeuAc linkages are evident in the MALDI-TOF glycan profiles in Figure 3 despite the expression of ST8Sia IV in THY, SPL, and LNG. However, this enzyme exhibits highly restricted specificity for N-glycans of neural cell-adhesion molecule (NCAM) and adds polymers of 2–8 linked sialic acids on these glycans. The absence of these structures is due either to a low or no expression of NCAM or to a low abundance and higher mass of polysialylated N-glycans. For O-linked glycans, each tissue expresses one or more of the ST3Gal I/II and the ST6GalNAc I–VI enzymes that form the NeuAc2–3Galß1–3 (NeuAc2–6)GalNAcThr/Ser sequence, the most common O-linked glycan. Most STs involved in the synthesis of glycoprotein glycans are able to transfer sialic acids into corresponding glycans of glycolipids. It is notable that BRN, unlike the other tissues, expresses high levels of the ST8Sial III and V enzymes responsible for the NeuAc2–8NeuAc linkages in glycolipids of the ganglioside series which occur in BRN at 10- to 100-fold higher levels than other tissues (Ando, 1983; Yoshida et al., 1995; Kono et al., 1996).
Fucosyltransferases comprise a family of ~13 members, which catalyze the synthesis of the Fuc1–2Gal (Fut1 and Fut2), Fuc1–3GlcNAc (Fut3–Fut7 and Fut9), Fuc1–4GlcNAc (Fut3), and Fuc1–6GlcNAc (Fut8) linkages. Of the remaining four, the Pofut1 and Pofut2 enzymes are involved in the formation of a novel O-linked glycan FucThr in peptide motifs found in epidermal growth factor domains, and no activity has yet been reported for Fut10 and Fut11 (Becker and Lowe, 2003). The Fuc1–6GlcNAc linkage is produced by Fut8 and is exclusively found in the core region of N-linked glycans attached to the GlcNAßAsn residue. As evident in Figure 3, N-linked glycans of most tissues are "core fucosylated" to some degree. Of the six tissues represented, the N-linked glycans of BRN and KID are predominately fucosylated (>90%), those of SPL and LNG are 50–75% fucosylated, and those of THY and LIV are <50% fucosylated. As summarized in Table II, expression of Fut8 is detected in all tissues except LIV. The highest levels of expression in BRN and KID correspond to the high degree of fucosylation in these tissues. Lack of detection of Fut8 expression in LIV is likely because of an expression level below the background signal, or that core fucosylation in this tissue is due instead to the existence of a yet unidentified Fut8 homologue (Miyoshi and Taniguchi, 2002).
Although the Fuc1–6GlcNAc sequence is unique to N-linked glycans, the other three fucose-containing linkages are found in terminal sequences of both N- and O-linked glycoprotein and glycolipid glycans. In contrast to the virtually ubiquitous presence of core fucosylation, the terminal Fuc1–2Gal sequence is not detected in any of the N-glycan profiles in Figure 3. Consistent with this observation, no expression was detected for the two fucosyltransferases capable of making this linkage, Fut1 and Fut2 (Table II). However, terminal sequences with the Fuc1–3(4)GlcNAc linkage are a dominant terminal sequence in BRN and KID (Figure 3). Of the enzymes capable of forming these sequences, Fut9 is the only one expressed in BRN and KID and is not expressed in any other tissue. Fut9 was previously demonstrated to be the enzyme responsible for the presence of Fuc1–3GlcNAc linkages in BRN (Nishihara et al., 1999, 2003). To determine whether this was also the case in KID, we performed immunohistochemistry analysis of KID sections from wild-type and Fut9 knockout mice, using an anti-stage-specific embryonic antigen 1 (SSEA-1) monoclonal antibody that recognizes the Galß1–4(Fuc1–3)GlcNAc (Lewis x) sequence as an antigenic epitope. As shown in Figure 4, proximal tubules of the wild-type KID are strongly stained with anti-SSEA-1, whereas no staining is observed in the Fut9 knockout mouse tissue, indicating a complete absence of the Fuc1–3GlcNAc. Western blotting of glycoproteins size fractionated on sodium dodecyl sulfate (SDS) gels, and immuno-thin layer chromatography of glycolipids also showed staining of wild-type glycoproteins and glycolipids, but not those from the Fut9 mouse. Thus, Fut9 appears to be the fucosyltransferase responsible for the synthesis of the Fuc1–3GlcNAc linkage in KID.
|
Taken together, the results show that evaluation of glycosyltransferase expression data in a focused microarray will provide an important source of meta data for investigating the regulation of glycan composition in mammalian tissues and cells.
Glycan-binding proteins
In addition to fulfilling structural roles, glycans function in many biological contexts involving recognition of glycan-encoded information by endogenous lectins or glycan-binding proteins. Animal lectins fall into multiple structural categories of which the C-type (calcium dependent) and I-type (sialic acid immunoglobulin superfamily lectins) are two of the most diverse (Taylor and Drickamer, 2003).
Examination of the expression data in Table III reveals multiple, contrasting patterns of expression. For example, a few of the C-type lectins appear ubiquitous or almost so. The observation that dectin-1, the beta-glucan receptor of macrophages, and the macrophage mannose receptor are observed to be present in both immune and non-immune tissues suggests that this pattern may be a hallmark of macrophage expression. Similarly, the broad distribution of P-selectin reflects its expression in endothelial cells in immune and non-immune tissues. In contrast, other genes are expressed in a very restricted set of tissues. For example, the surfactant proteins are only present in LNG, brevican only in BRN, and the mannose-binding proteins and galactose-binding clearance receptors, such as the hepatic asialoglycoprotein receptor and the Kupffer cell receptor, are LIV specific.
|
A group of genes restricted to the immune tissues are evident. L-selectin is constitutively and almost exclusively found on lymphocytes (Vestweber and Blanks, 1999) and is thus present in tissues rich in immune cells, as are most of the proteins included in subcategory five of the C-type lectins, which are associated with T cells (CD69), B cells (CD72), or natural killer (NK) cells (Robinson et al., 1993; Brown et al., 1997). The various mannose- and galactose-binding type II receptors in subcategory two are the least well understood of the C-type lectins, and their expression data provide new insights. For example, the macrophage asialoglycoprotein (galactose) receptor is known to be expressed on peritoneal and tumoricidal subsets of macrophages and on dendritic cells (Tsuiji et al., 2002), but the high level in lymph nodes might suggest expression in further cell types. On the other hand, the IgE Fc receptor (CD23) is expressed primarily on lymphocytes (Lüdin et al., 1987), so expression restricted to the immune tissues is expected. Dendritic cell (DC) immunoreceptor has been described as a macrophage protein, as well as being on monocytes, B cells and granulocytes (Bates et al., 1999), and the data show the general distribution typical of macrophage expression. In contrast, mincle, which is usually described as a macrophage protein (Matsumoto et al., 1999) is restricted to immune tissues, which is not really expected and may be indicative of expression in another cell type, such as dendritic cells. The main observation about the members of the DC-specific ICAM3-grabbing non-integrin (DC-SIGN) family is that the expression patterns for different family members seem to differ, suggesting that these proteins are involved in different functions or at least in the same function on different cell types. The finding of high levels of SIGN-R1 in SPL and lymph node corroborates recent descriptions of expression specifically in marginal zone macrophages of these two tissues (Geijtenbeek et al., 2002).
The Siglecs group includes members of the Ig superfamily that bind sialic acid and are known to be expressed predominately by hematopoietic cells. Siglecs can bind either cis (on the same cell) or trans sialic acid and therefore are believed to regulate cell interaction. Moreover Siglecs possess cytoplasmic immuno-receptor tyrosine-based inhibitory motif (ITIM) and can consequently be involved in regulating cell activation (Crocker and Varki, 2001; Crocker, 2002; Nitschke and Tsubata, 2004). Consistent with these features, and conversely to the C-type lectins, all but one of the Siglecs detected by the array were found to be predominately expressed in the immune tissues. In particular, CD22 (Siglec-2; predominately expressed B cells), Siglec-E (neutrophils, NK cells, and dendritic cells), and Siglec-10 (B cells, eosinophils) are predominately expressed in BM, lymph nodes, and SPL, and Siglec-F, selectively expressed in eosinophils, is expressed in BM, lymph nodes, and LNG (Crocker, 2002; Aizawa et al., 2003; Zhang et al., 2004). The one exception is Siglec-4, or myelin-associated glycoprotein, which is expressed in glial cells of the central nervous system and, accordingly, was found to be restricted to BRN (Crocker, 2002).
Proteoglycans
Proteoglycans consist of one or more glycosaminoglycan chains (heparan sulfate, chondroitin sulfate, or dermatan sulfate) attached to serine residues within a limited set of core proteins. Some proteoglycans are secreted and deposited in basement membranes (agrin and perlecan), whereas others make up the extracellular matrix surrounding connective tissue (e.g., the aggrecan family, the collagen proteoglycans, IX and XIV, as well as the small leucine-rich proteoglycans, decorin, biglycan, lumican, fibromodulin, and others). Many membrane-associated proteoglycans exist as well, such as syndecans 1–4, the GPI-anchored glypicans 1–6, a CD44 variant called epican, and others not represented on the array (e.g., NG2). Intracellular proteoglycans include serglycin, a proteoglycan found in storage granules, and invariant chain (CD74).
Glypican-1 tends to be expressed widely, whereas glypican-2 is expressed in the developing nervous system and disappears upon completion of cell migration and axon outgrowth. Gypican-3 (OCI-5) is expressed during development in several tissues, but in the adult its expression is nearly restricted to the LNG. Glypican-4 (K-glypican) is highly expressed in the mouse KID and at low levels in adult BRN and SPL. Glypican-5 is highly expressed in the BRN in adult animals. Glypican-6 is expressed most abundantly in the ovary, LIV, and KID, with lower levels of expression detected in other adult tissues. The microarray analysis summarized in Table IV, in general, confirms these findings (De Cat and David, 2001; Filmus and Selleck, 2001; Fransson, 2003).
|
As expected, the small leucine rich proteoglycans are expressed in nonlymphoid tissues, but the high level of expression in THY, lymph node, and SPL suggests that their expression may be tied to other functions. A similar argument can be made for collagen type XIV, which tends to be expressed in tissues containing fibrillar collagens. Decorin exhibits growth inhibitory activity, possibly by way of binding transforming growth factor-ß (TGF-ß), and its high level of expression in THY and lymph node suggests a potential role in immune cell maturation (Castagnola et al., 1992; Hocking et al., 1998; Iozzo, 1998).
Not surprisingly, serglycin is expressed throughout the lymphoid tissues, most likely as a granular proteoglycan present in virtually all hematopoietic cells. CD44 (epican) and invariant chain are similarly expressed at high levels in lymphoid tissue, where they play a role in B-cell maturation. Syndecans are conspicuously missing from lymphoid tissues and cells, which may reflect their primary role in epithelial cells (Kolset and Gallagher, 1990; Schofield et al., 1999; Couchman, 2003).
The heparan sulfate bearing proteoglycans play critical roles in embryonic development and physiology of adult animals, but their role in the immune system is unclear. Because many growth factors, chemokines, and cytokines bind to heparan sulfate, they may play a role in matrix immobilization of these ligands or they may play direct roles in signaling reactions. The chondroitin sulfate bearing proteoglycans play important roles in matrix deposition and tissue architecture. Other roles in the immune system are not well studied. The microarray data suggest that additional studies are needed to understand how these proteoglycans participate in lymphoid (and myeloid) development, trafficking, and immune responses (Lander and Selleck, 2000; Perrimon and Bernfield, 2000; Esko and Lindahl, 2001; Esko and Selleck, 2002).
|
Discussion |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsReferences |
|---|
This study describes the establishment of a gene expression array, built with the Affymetrix platform, representing 436 human and 285 murine glycogenes and its use to investigate their expression patterns across nine murine tissues. Results cover the expression of many genes for which data in the literature are nonexistent, scarce, or inconsistent. Observations derived from this analysis provide insights on the extent of transcriptional regulation of glycosylation. Since this study was completed, CFG has developed an updated glycogene-chip v2 (Smith et al., 2005
), with 532 human and 470 murine glycogenes, that is available for use by request, and the gene list available online (http://www.functionalglycomics.org). Despite the low levels at which most glycogenes are expressed, 90% were scored present in at least one tissue. On the other hand, only 15% were present in all tissues examined, highlighting the specificity of glycogene expression. Notably, the glycosyltransferases involved in synthesis of core structures of N- and O-glycans were ubiquitously expressed, whereas expression of those involved in terminal structures was highly variable. Hierarchical clustering analysis further illustrated this finding, showing that tissues with related biological function cluster together based on their glycogene expression profiles. This was true for glycogenes belonging to different classes, such as the glycosyltransferases, the proteoglycans, and the glycan-binding proteins. In particular, two of the major families of glycan-binding proteins, C-type lectins, and Siglecs, are predominately expressed in the immune tissues, consistent with the increasing evidence for their important roles in both innate and acquired immune function.
As described in Results, the validity of the gene expression data obtained with this chip is corroborated by the fact that the many of the previously described gene-expression patterns have been confirmed by the array data. In some of the cases where a gene was not detected in a tissue where its expression is known, we could identify faults in the probe design mostly because of lacking sequence information of the 3' untranslated region, which can be extremely long for some genes, particularly glycosyltransferases. Because this problem was anticipated, an effort was made to include well-characterized expressed sequence tags (EST) or proprietary sequences provided by the CFG investigators. We believe that redesign of probe sets can resolve most of the problems of poor sensitivity. For genes in glycogene-chip v1 that gave no signal, 90% were detected following redesign of the probe sets in glycogene-chip v2 (S. Head and T. Gilmartin, unpublished data).
From a general perspective, the significance of results from microarray data, as for other gene expression data, is most powerful when used in conjunction with other biochemical data, as demonstrated for glycosyltransferase expression and glycan profiling in this report. Thus, genes involved in the synthesis of the core of ubiquitous N-glycans are coexpressed at similar levels across all tissues. By contrast, glycosyltransferases that elaborate terminal structures in N- and O-linked glycoproteins and in glycolipids are expressed in a highly tissue-specific manner, resulting in terminal glycosylation patterns as seen in the MALDI profiles for N-glycans. Use of this approach for linking glycosyltransferase expression to glycan structure was nicely exemplified in this study by identifying Fut9 as the enzyme responsible for the synthesis of Lewis x structures in KID. Such findings are consistent with the idea that terminal glycan sequences modulate the biological function of the protein to which they are attached. Therefore, the data support the increasing anecdotal evidence that differences between the glycome of diverse organs or cell types, or within the same cell type at different physiological states, largely depend on the concerted variation of glycosyltransferase gene expression levels.
The low expression levels of glycosyltransferases (Weinstein et al., 1982; Ju et al., 2002), together with the fact that more than one enzyme can be responsible for the synthesis of the same glycan structure, has impaired a systematic analysis of glycoconjugates expression regulation. Microarray technology allows a big step forward in this direction. Enormous potential for discovery exists by performing meta-analysis of data sets for gene expression and glycan profiling by mass spectrophotometry techniques conducted on the same tissues and cell populations.
|
Materials and methods |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsReferences |
|---|
Glycogene-chip v1 design
The glycogene-chip v1 oligonucleotide array is a custom Affymetrix GeneChip designed for the CFG (http://www.functionalglycomics.org). The array contains probe sets targeted to 436 human and 285 murine glycogenes, including glycosyltransferases, glycan-binding proteins, proteoglycans, and other genes involved in glycan synthesis and degradation. In addition, the chip contains probe sets for several other classes of genes including chemokines, cytokines, growth factors, interleukins, adhesion molecules, and associated receptors. A set of probes for housekeeping genes was also included, as well as for 45 "spiked-in" genes from Escherichia coli, Bacillus subtilis, and bacteriophage for quality control. In all, there are 624 mouse genes (752 probe sets) and 915 human genes (1017 probe sets) represented. The list of the glycogenes was compiled using information available on public databases, such as UniGene (http://www.ncbi.nlm.nih.gov/entrez/query.fcgi?db=unigene) and CAZy (http://afmb.cnrs-mrs.fr/~cazy/CAZY/index.html), or using unpublished information provided by participating investigators of the CFG. The participating investigators also participated in extensive annotation of the list. The highly annotated gene list is available at http://www.functionalglycomics.org. This chip uses the standard Affymetrix gene expression detection system, described by Lockhart et al. (1996)
. Each transcript is targeted by one or more probe sets consisting of 11 probe pairs, each made of a perfect match (PM) and a mismatch (MM) 25 base oligonucleotides. The chip contains two independent groups of probe sets, one for human and the other for mouse transcripts. The human probe sets were chosen, whenever possible, from the commercially available U133A,B GeneChip set, otherwise they were newly designed. Only 374 of the 436 human glycogenes are found or annotated on the Affymetrix U133A,B chip set, highlighting the incompleteness of this commercial array. In contrast, all mouse gene probe sets were custom designed specifically for the glycogene-chip v1. No attempt was made to screen for cross-hybridization between mouse and human species during the probe-design process.
Mice and tissue samples preparation
Mouse tissues comprising of LIV, BRN, KID, SPL, THY, lymph nodes, testis, LNG, and femurs for BM preparation were harvested from 6 to 8 weeks old C57BL/6 male mice obtained from the Scripps Research Institute (La Jolla, CA) custom breeding core and sacrificed by cervical dislocation. The organs were snapfrozen immediately after harvesting. For BM cell-suspension preparation, the tissue was grinded between two frosted glass slides in Roswell Park Memorial Institute (RPMI) medium, and the obtained suspension clarified by filtering through cotton-plugged pipette before lysing erythrocytes by incubating at room temperature for 5 min in a solution of 150 mM ammonium chloride, 10 mM potassium carbonate, 0.1 mM ethylene diamine tetraacetic acid (EDTA), pH 7.2. All samples were stored at –80°C.
RNA preparation, labeling, and chip hybridization
Total RNA was extracted with Trizol, according to the manufacturer’s suggested protocol (Invitrogen, Carlsbad, CA), treated with Ambion DNA-free DNase treatment (Ambion, Austin, TX), and purified with Qiagen Rneasy columns using the cleanup protocol (Qiagen, Valencia, CA). The quality of the samples was checked with an Agilent Bioanalyzer (Agilent Technologies, Palo Alto, CA). RNA from three independent preparations, each composed of three mice, was amplified and biotin labeled using the Bioarray High Yield RNA transcript labeling kit (Enzo Life Sciences, Farmingdale, NY). Hybridization and scanning of the glycogene-chip v1 were performed according to the Affymetrix’s recommended protocols (Affymetrix, Santa Clara, CA).
Microarray data analysis
Present (P) and absent (A) absolute calls were determined with the MicroArray Suite (MAS) 5.0 Affymetrix algorithm. The glycogenes were subsequently considered P in one tissue if they had been assigned a P call in at least two of the three replicate samples. All marginal (M) calls were interpreted as A.
At present, there is no consensus on procedures for analyzing microarray data. We therefore evaluated several available methods seeking the most suitable for the glycogene-chip v1. To choose the signal generating algorithm, we evaluated MAS 5.0 (Affymetrix, Santa Clara, CA), dChip (Li and Wong, 2001), and robust multiarray analysis (RMA) (Irizarry et al., 2003). To test the accuracy of these methods, we designed an experiment in which labeled RNAs from two mouse tissues, KID, and SPL, were prepared and mixed in a range of ratios and applied to replicate chips for analysis. The RNA preparations included 100% KID, 100% SPL, 75% KID 25% SPL, 50% KID 50% SPL, and 25% KID 75% SPL. The design of this experiment was based on the assumption that transcript levels will change linearly while going from 100% of one tissue to the other. We observed a higher number of genes with Pearson’s correlation coefficient closer to 1 when using dChip and RMA compared with MAS 5.0 (Figure 5). This implies that by using the dChip and RMA signal algorithms, more of the variation between signal values could be caused by the biological variation (systematic mixing of the tissue types) than by technical variation. We decided to use the dChip PM-only model because of the general ease of use of the software at the time the analysis was performed. The dChip algorithm was therefore used to generate expression signal intensities and to normalize the data set against the median-intensity array using the invariant-set strategy (Li and Wong, 2001). A lack of consensus also exists about the applicability of normalization methods for certain types of arrays, including focused arrays with potentially large proportions of genes exhibiting differential expression and where general statistical assumptions may be inaccurate (Hill et al., 2001; Hoffmann et al., 2002; Irizarry et al., 2003). Because the glycogene v1 array is a small focused chip, our initial concern was that the invariant-set normalization method employed by dChip might prove inappropriate. Therefore, to assess the utility of this normalization strategy, we examined the signals generated by the prelabeled "spike-in" controls which are routinely added to GeneChip hybridization cocktails (Eukaryotic hybridization controls, Affymetrix, Santa Clara, CA) in the nine different wild-type tissues surveyed in this study. We tested the spike control signals (BioB, BioC, BioD, and Cre) and determined, by single factor analysis of variance (ANOVA), that no significant difference existed across the nine wild-type tissues (p > 0.05). We therefore concluded that this normalization method is appropriate for use with this chip.
|
The entire data set has been submitted to GEO repository under accession number GSE2781
[NCBI GEO]
. Genes differentially expressed across groups of replicate chips were identified by significance of microarray analysis (SAM) (Tusher et al., 2001). Following determination of the test statistic (q value), the software performs many permutations that facilitates calculation of the false discovery rate (FDR), which is defined as the proportion of false positives among the significant genes (Broberg, 2003). Genes were identified as having a significant change in expression if their associated FDR was <1%. Unsupervised hierarchical clustering by sample was performed with Biometric Research Branch (BRB) ArrayTools (http://linus.nci.nih.gov/BRB-ArrayTools.html) using the default settings for correlation and linkage.
Chip-performance assessment
We first evaluated the performance of the glycogene-chip v1 against the existing commercial array Affymetrix U133A, in particular to determine the effects of the glycogene-chip size on individual gene expression measurements. To this end, we examined a subset of 835 human probe sets with identical probe sequences common to both the glycogene-chip v1 (1814 probe sets) and the Affymetrix U133A chip (22,000+ probe sets). We prepared RNA from the LS180 (human colon carcinoma) and U937 (human lymphoma) cell lines and hybridized each RNA preparation to three chips of each type. We compared the expression of the 835 probes subset across the two chip types for each cell line, using the MAS 5.0 software’s algorithm for P/A determination. The average percent P was 39.2 (±2.5) and 35.0 (±2.5) for the glycogene-chip v1 and the U133A chip, respectively. In addition, the chip background for both chip types was very similar (within 10%). These results suggest that, for the subset of genes found on both chips, the probe sets on the smaller glycogene chip detect a higher percentage of transcripts as "P" and perform at least as well as their counterparts on the large format U133A chip.
We then addressed experimental reproducibility and measured nonbiological variability resulting from the sample preparation and labeling procedures. The experiment consisted of 36 separate array hybridizations using mouse BRN RNA. In brief, RNA was prepared from four aliquots of a single homogenized mouse tissue sample by two different technicians. Three labeling reactions were performed for each RNA sample, and each labeling reaction was hybridized to three chips. Of the 33 chips that met our quality control specifications, we measured a mean coefficient of determination (r2) of 0.951 and a standard deviation of 0.060. In addition, hierarchical clustering analysis failed to elucidate a correlation associated with the technical parameters being tested (data not shown). These data suggest that the nonbiological variability associated with sample RNA extraction, labeling, and hybridization is relatively low and, in our laboratory, has little systematic impact on the resulting microarray data. Based on these results, all subsequent experiments were performed with three independently prepared RNA preparations from each tissue, and a single labeling was used to probe a single chip for each RNA sample.
N-Glycan profiling
N-Glycans were isolated from trypsinized detergent extracts of homogenized organs by peptide : N-glycanase F (PNGase F), permethylated by using the sodium hydroxide procedure, and purified on a Sep-Pak C 18 cartridge, as previously described (Sutton-Smith et al., 2000). MALDI-MS instrumentation was the same, as previously described (Kui Wong et al., 2003). BRN, KID, LIV, LNG, and SPL were analyzed at M-Scan Inc (West Chester, PA), and THY was analyzed at Imperial College London (London, UK). Derivatized carbohydrate samples were dissolved in methanol/water 8:2 (v/v) and mixed in a 1:1 ratio with 10 mg/mL 2,5-dihydroxybenzoic acid in 80:20 (v/v) methanol/water. About 1.5 µL aliquots were spotted onto a 100-well sample plate. Angiotensin I, adrenocorticotropic hormone (ACTH) fragment 1–17, ACTH fragment 18–39, and ACTH fragment 7–38, were used for external calibration.
Immunohistochemical analysis of Lewis x expression in murine KID
Expression of the 3-fucosyl-N-acetyllactosamine carbohydrate epitope was examined immunohistochemically with a mouse monoclonal antibody, SSEA-1 (IgM, Developmental Studies Hybridoma Bank at the University of Iowa, Iowa City), at dilution of 1/500, using the method reported previously (Kudo et al., 2004). KID tissues of wild-type and Fut9 knockout mice were fixed in 4% paraformaldehyde/phosphate-buffer saline (–) (w/v) and then embedded in paraffin. Immunostaining was performed using the streptavidin–biotin complex method (Elite ABC kit; Vectastain; Vector Laboratories, Burlingame, CA) in deparaffinized sections. The peroxidase reaction was visualized with 3,3'-diaminobenzidine, and nuclei were counterstained with hematoxylin.
|
Acknowledgments |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsReferences |
|---|
We are indebted to Margarida Amado, Henrik Clausen, Bret Crawford, Paul Crocker, Richard Cummings, Minoru Fukuda, Hakon Leffler, Bernard Henrissat, Carlos Hershberg, John Lowe, Risto Renkonen, Steve Rosen, Lianchun Wang, and investigators of the Consortium for Functional Glycomics for their contributions during chip design and stimulating scientific discussion. We thank Brian Collins and Karin Norgaard-Sumnicht for help with sample treatment, Caroline Lanigan for support with the statistical analysis, and Anna Tran-Crie for assistance with the manuscript preparation. This work was supported by National Institute of General Medical Sciences grants GM62116 to the Consortium for Functional Glycomics and GM33063 to J.D.E. and AI50143 to J.P., by Biotechnology and Biological Sciences Research Council (BBSRC) and Wellcome Trust grants to A.D., and support to H.N. from the New Energy and Industrial Technology Development Organization (NEDO) of Japan as part of the R & D Project of Industrial Science and Technology Frontier Program. A.D. is a BBSRC Professorial Research Fellow. E.M.C. was partially funded by Nestlé Research Centre, Lausanne, Switzerland.
|
Abbreviations |
|---|
ACTH, adrenocorticotropic hormone; BM, bone marrow; BRN, brain; CFG, Consortium for Functional Glycomics; DC, dendritic cell; DC-SIGN, DC-specific ICAM3-grabbing non-integrin; Fut9, fucosyltransferase 9; KID, kidney; LIV, liver; LNG, lung; MALDI-TOF MS, matrix-assisted laser desorption ionization-time of flight mass spectrometry; MAS, MicroArray Suite; NK, natural killer; ppGalNAcT, polypeptide N-acetylgalactosaminyl transferase; SPL, spleen; SSEA-1, stage-specific embryonic antigen 1; ST, sialyltransferase; TES, testes; THY, thymus
|
References |
|---|
|
TopAbstractIntroductionResultsDiscussionMaterials and methodsAcknowledgmentsReferences |
|---|
Aizawa, H., Zimmermann, N., Carrigan, P.E., Lee, J.J., Rothenberg, M.E., and Bochner, B.S. (2003) Molecular analysis of human Siglec-8 orthologs relevant to mouse eosinophils: identification of mouse orthologs of Siglec-5 (mSiglec-F) and Siglec-10 (mSiglec-G). Genomics, 82, 521–530.[CrossRef][Web of Science][Medline]
Ando, S. (1983) Gangliosides in the nervous system. Neurochem. Int., 5, 507–537.[CrossRef][Web of Science]
Apweiler, R., Hermjakob, H., and Sharon, N. (1999) On the frequency of protein glycosylation, as deduced from analysis of the SWISS-PROT database. Biochim. Biophys. Acta, 1473, 4–8.[Medline]
Bates, E.E.M., Fournier, N., Garcia, E., Valladeau, J., Durand, I., Pin, J.-J., Zurawski, S.M., Patel, S., Abrams, J.S., Lebecque, S., and others. (1999) APCs express DCIR, a novel C-type lectin surface receptor containing and immunoreceptor tyrosine-based inhibitory motif. J. Immunol., 163, 1973–1983.
Becker, D.J. and Lowe, J.B. (2003) Fucose: biosynthesis and biological function in mammals. Glycobiology, 13, 41R–53R.
Bennett, E.P., Hassan, H., and Clausen, H. (1996) cDNA cloning and expression of a novel human UDP-N-acetyl-alpha-D-galactosamine. Polypeptide N-acetylgalactosaminyltransferase, GalNAc-t3. J. Biol. Chem., 271, 17006–17012.
Bosques, C.J., Tschampel, S.M., Woods, R.J., and Imperiali, B. (2004) Effects of glycosylation on peptide conformation: a synergistic experimental and computational study. J. Am. Chem. Soc., 126, 8421–8425.[CrossRef][Web of Science][Medline]
Broberg, P. (2003) Statistical methods for ranking differentially expressed genes. Genome Biol., 4, R41.
Brown, M.G., Scalzo, A.A., Matsumoto, K., and Yokoyama, W.M. (1997) The natural killer gene complex: a genetic basis for understanding natural killer cell function and innate immunity. Immunol. Rev., 155, 53–65.[CrossRef][Web of Science][Medline]
Brown, J.R., Fuster, M.M., Whisenant, T., and Esko, J.D. (2003) Expression patterns of alpha 2,3-sialyltransferases and alpha 1,3-fucosyltransferases determine the mode of sialyl Lewis X inhibition by disaccharide decoys. J. Biol. Chem., 278, 23352–23359.
Castagnola, P., Tavella, S., Gerecke, D.R., Dublet, B., Gordon, M.K., Seyer, J., Cancedda, R., van der Rest, M., and Olsen, B.R. (1992) Tissue-specific expression of type XIV collagen – a member of the FACIT class of collagens. Eur. J. Cell Biol., 59, 340–347.[Web of Science][Medline]
Colmenares, M., Corbi, A.L., Turco, S.J., and Rivas, L. (2004) The dendritic cell receptor DC-SIGN discriminates among species and life cycle forms of Leishmania. J. Immunol., 172, 1186–1190.
Cooper, C.A., Joshi, H.J., Harrison, M.J., Wilkins, M.R., and Packer, N.H. (2003) GlycoSuiteDB: a curated relational database of glycoprotein glycan structures and their biological sources. 2003 update. Nucleic Acids Res., 31, 511–513.
Couchman, J.R. (2003) Syndecans: proteoglycan regulators of cell-surface microdomains? Nat. Rev. Mol. Cell Biol., 4, 926–937.[CrossRef][Web of Science][Medline]
Crocker, P.R. (2002) Siglecs: sialic-acid-binding immunoglobulin-like lectins in cell-cell interactions and signalling. Curr. Opin. Struct. Biol., 12, 609–615.[CrossRef][Web of Science][Medline]
Crocker, P.R. and Varki, A. (2001) Siglecs, sialic acids and innate immunity. Trends Immunol., 22, 337–342.[CrossRef][Web of Science][Medline]
De Cat, B. and David, G. (2001) Developmental roles of the glypicans. Semin. Cell Dev. Biol., 12, 117–125.[CrossRef][Web of Science][Medline]
Esko, J.D. and Lindahl, U. (2001) Molecular diversity of heparan sulfate. J. Clin. Invest., 108, 169–173.[CrossRef][Web of Science][Medline]
Esko, J.D. and Selleck, S.B. (2002) Order out of chaos: assembly of ligand binding sites in heparan sulfate. Annu. Rev. Biochem., 71, 435–471.[CrossRef][Web of Science][Medline]
Filmus, J. and Selleck, S.B. (2001) Glypicans: proteoglycans with a surprise. J. Clin. Invest., 108, 497–501.[CrossRef][Web of Science][Medline]
Fransson, L.A. (2003) Glypicans. Int. J. Biochem. Cell Biol., 35, 125–129.[CrossRef][Web of Science][Medline]
Fukui, S., Feizi, T., Galustian, C., Lawson, A.M., and Chai, W. (2002) Oligosaccharide microarrays for high-throughput detection and specificity assignments of carbohydrate-protein interactions. Nat. Biotechnol., 20, 1011–1017.[CrossRef][Web of Science][Medline]
Geijtenbeek, T.B.H., Groot, P.C., Nolte, M.A., van Vliet, S.J., Gangaram-Panday, S.T., van Duijnhoven, G.C.F., Kraal, G., van Oosterhout, A.J.M., and van Kooyk, Y. (2002) Marginal zone macrophages express a murine homologue of DC-SIGN that captures blood-borne antigens in vivo. Blood, 100, 2908–2916.
Harduin-Lepers, A., Vallejo-Ruiz, V., Krzewinski-Recchi, M.A., Samyn-Petit, B., Julien, S., and Delannoy, P. (2001) The human sialyltransferase family. Biochimie, 83, 727–737.[Medline]
Helenius, A. and Aebi, M. (2004) Roles of N-linked glycans in the endoplasmic reticulum. Annu. Rev. Biochem., 73, 1019–1049.[CrossRef][Web of Science][Medline]
Henrissat, B. and Bairoch, A. (1996) Updating the sequence-based classification of glycosyl hydrolases. Biochem. J., 316, 695–696.[Web of Science][Medline]
Hill, A.A., Brown, E.L., Whitley, M.Z., Tucker-Kellogg, G., Hunter, C.P., and Slonim, D.K. (2001) Evaluation of normalization procedures for oligonucleotide array data based on spiked cRNA controls. Genome. Biol., 2, RESEARCH0055.
Hocking, A.M., Shinomura, T., and McQuillan, D.J. (1998) Leucine-rich repeat glycoproteins of the extracellular matrix. Matrix Biol., 17, 1–19.[CrossRef][Web of Science][Medline]
Hoffmann, R., Seidl, T., and Dugas, M. (2002) Profound effect of normalization on detection of differentially expressed genes in oligonucleotide microarray data analysis. Genome Biol., 3, RESEARCH0033.
Iozzo, R.V. (1998) Matrix proteoglycans: from molecular design to cellular function. Annu. Rev. Biochem., 67, 609–652.[CrossRef][Web of Science][Medline]
Irizarry, R.A., Bolstad, B.M., Collin, F., Cope, L.M., Hobbs, B., and Speed, T.P. (2003) Summaries of Affymetrix GeneChip probe level data. Nucleic Acids Res., 31, e15.
Ju, T., Cummings, R.D., and Canfield, W.M. (2002) Purification, characterization, and subunit structure of rat core 1 beta1,3-galactosyltransferase. J. Biol. Chem., 277, 169–177.
Kanehisa, M., Goto, S., Kawashima, S., Okuno, Y., and Hattori, M. (2004) The KEGG resource for deciphering the genome. Nucleic Acids Res., 32, D277–D280.
Karlsson, K.A. (1995) Microbial recognition of target-cell glycoconjugates. Curr. Opin. Struct. Biol., 5, 622–635.[CrossRef][Web of Science][Medline]
Kikuchi, N., Kwon, Y.D., Gotoh, M., and Narimatsu, H. (2003) Comparison of glycosyltransferase families using the profile hidden Markov model. Biochem. Biophys. Res. Commun., 310, 574–579.[CrossRef][Web of Science][Medline]
Kolset, S.O. and Gallagher, J.T. (1990) Proteoglycans in haemopoietic cells. Biochim. Biophys. Acta, 1032, 191–211.[Medline]
Kono, M., Yoshida, Y., Kojima, N., and Tsuji, S. (1996) Molecular cloning and expression of a fifth type of alpha2,8-sialyltransferase (ST8Sia V). Its substrate specificity is similar to that of SAT-V/III, which synthesize GD1c, GT1a, GQ1b and GT3. J. Biol. Chem., 271, 29366–29371.
Kudo, T., Kaneko, M., Iwasaki, H., Togayachi, A., Nishihara, S., Abe, K., and Narimatsu, H. (2004) Normal embryonic and germ cell development in mice lacking alpha 1,3-fucosyltransferase IX (Fut9) which show disappearance of stage-specific embryonic antigen 1. Mol. Cell Biol., 24, 4221–4228.
Kui Wong, N., Easton, R.L., Panico, M., Sutton-Smith, M., Morrison, J.C., Lattanzio, F.A., Morris, H.R., Clark, G.F., Dell, A., and Patankar, M.S. (2003) Characterization of the oligosaccharides associated with the human ovarian tumor marker CA125. J. Biol. Chem., 278, 28619–28634.
Lander, A.D. and Selleck, S.B. (2000) The elusive functions of proteoglycans: in vivo veritas. J. Cell Biol., 148, 227–232.
Li, C. and Wong, W.H. (2001) Model-based analysis of oligonucleotide arrays: expression index computation and outlier detection. Proc. Natl. Acad. Sci. U. S. A., 98, 31–36.
Lockhart, D.J., Dong, H., Byrne, M.C., Follettie, M.T., Gallo, M.V., Chee, M.S., Mittmann, M., Wang, C., Kobayashi, M., Horton, H., and Brown, E.L. (1996) Expression monitoring by hybridization to high-density oligonucleotide arrays. Nat. Biotechnol., 14, 1675–1680.[CrossRef][Web of Science][Medline]
Lowe, J.B. (2002) Glycosylation in the control of selectin counter-receptor structure and function. Immunol. Rev., 186, 19–36.[CrossRef][Web of Science][Medline]
Lüdin, C., Hofstetter, H., Sargati, M., Levy, C.A., Suter, U., Alaimo, D., Kilchherr, E., Forst, H., and Delespesse, G. (1987) Cloning and expression of the cDNA coding for a human lymphocyte IgE receptor. EMBO J., 6, 109–114.[Web of Science][Medline]
Lutteke, T., Frank, M., and von der Lieth, C.W. (2004) Data mining the protein data bank: automatic detection and assignment of carbohydrate structures. Carbohydr. Res., 339, 1015–1020.[CrossRef][Web of Science][Medline]
Matsumoto, M., Tanaka, T., Kaisho, T., Sanjo, H., Copeland, N.G., Gilbert, D.J., Jenkins, N.A., and Akira, S. (1999) A novel LPS-inducible C-type lectin is a transcriptional target of NF-IL6 in macrophages. J. Immmunol., 163, 5039–5048.
Miyoshi, E. and Taniguchi, N. (2002) Alpha6-fucosyltransferase 8 (FUT8). In Taniguchi, N., Honke, K., and Fukuda, M. (eds), Handbook of Glycosyltransferases and Related Genes. Springer, Tokyo, pp. 259–263.
Nishihara, S., Iwasaki, H., Kaneko, M., Tawada, A., Ito, M., and Narimatsu, H. (1999) Alpha1,3-fucosyltransferase 9 (FUT9; Fuc-TIX) preferentially fucosylates the distal GlcNAc residue of polylactosamine chain while the other four alpha1,3FUT members preferentially fucosylate the inner GlcNAc residue. FEBS Lett., 462, 289–294.[CrossRef][Web of Science][Medline]
Nishihara, S., Iwasaki, H., Nakajima, K., Togayachi, A., Ikehara, Y., Kudo, T., Kushi, Y., Furuya, A., Shitara, K., and Narimatsu, H. (2003) Alpha1,3-fucosyltransferase IX (Fut9) determines Lewis X expression in brain. Glycobiology, 13, 445–455.
Nitschke, L. and Tsubata, T. (2004) Molecular interactions regulate BCR signal inhibition by CD22 and CD72. Trends Immunol., 25, 543–550.[CrossRef][Web of Science][Medline]
Perrimon, N. and Bernfield, M. (2000) Specificities of heparan sulphate proteoglycans in developmental processes. Nature, 404, 725–728.[CrossRef][Medline]
Rabinovich, G.A., Baum, L.G., Tinari, N., Paganelli, R., Natoli, C., Liu, F.T., and Iacobelli, S. (2002) Galectins and their ligands: amplifiers, silencers or tuners of the inflammatory response? Trends Immunol., 23, 313–320.[CrossRef][Web of Science][Medline]
Robinson, W.H., Landolfi, M.M., Schafer, H., and Parnes, J.R. (1993) Biochemical identity of the mouse Ly-19.2 and Ly-32.2 alloantigens with the B cell differentiation antigen Lyb-2/CD72. J. Immunol., 151, 4764–4772.[Abstract]
Schofield, K.P., Gallagher, J.T., and David, G. (1999) Expression of proteoglycan core proteins in human bone marrow stroma. Biochem. J., 343, 663–668.[CrossRef][Web of Science][Medline]
Smith, A.E. and Helenius, A. (2004) How viruses enter animal cells. Science, 304, 237–242.
Smith, F.I., Qu, Q., Hong, S.J., Kim, K.S., Gilmartin, T.J., and Head, S.R. (2005) Gene expression profiling of mouse postnatal cerebellar development using oligonucleotide microarrays designed to detect differences in glycoconjugate expression. Gene Expr. Patterns, 5, 740–749.[CrossRef][Medline]
Spiro, R.G. (2002) Protein glycosylation: nature, distribution, enzymatic formation, and disease implications of glycopeptide bonds. Glycobiology, 12, 43R–56R.
Sutton-Smith, M., Morris, H.R., and Dell, A. (2000) A rapid mass spectrometric strategy suitable for the investigation of glycan alterations in knockout mice. Tetrahedron Asymmetry, 11, 363–369.[CrossRef][Web of Science]
Sutton-Smith, M., Morris, H.R., Grewal, P.K., Hewitt, J.E., Bittner, R.E., Goldin, E., Schiffmann, R., and Dell, A. (2002) MS screening strategies: investigating the glycomes of knockout and myodystrophic mice and leukodystrophic human brains. Biochem. Soc. Symp., 105–115.
Taylor, M.E. and Drickamer, K. (2003) Introduction to Glycobiology. Oxford University Press, Oxford.
Ten Hagen, K.G., Fritz, T.A., and Tabak, L.A. (2003) All in the family: the UDP-GalNAc: polypeptide N-acetylgalactosaminyltransferases. Glycobiology, 13, 1R–16R.
Tsuiji, M., Fujimori, M., Ohashi, Y., Higashi, N., Onami, T.M., Hedric, S.M., and Irimura, T. (2002) Molecular cloning and characterization of a novel mouse macrophage C-type lectin, mMGL2, which has a distinct carbohydrate specificity from mMGL1. J. Biol. Chem., 277, 28892–28901.
Tusher, V.G., Tibshirani, R., and Chu, G. (2001) Significance analysis of microarrays applied to the ionizing radiation response. Proc. Natl. Acad. Sci. U. S. A., 98, 5116–5121.
Underhill, D.M. (2003) Toll-like receptors: networking for success. Eur. J. Immunol., 33, 1767–1775.[CrossRef][Web of Science][Medline]
Varki, A. (1999) Glycosphingolipids. In Varki, A., Cummings, R., Esko, J., Freeze, H., Hart, G., and Marth, J. (eds), Essentials of Glycobiology. Cold Spring Harbor Laboratory Press, Cold Spring Harbor, New York, pp. 115–129.
Vestweber, D. and Blanks, J.E. (1999) Mechanisms that regulate the function of the selectins and their ligands. Physiol. Rev., 79, 181–213.
von der Lieth, C.W., Bohne-Lang, A., Lohmann, K.K., and Frank, M. (2004) Bioinformatics for glycomics: status, methods, requirements and perspectives. Brief Bioinform., 5, 164–178.
Wandall, H.H., Hassan, H., Mirgorodskaya, E., Kristensen, A.K., Roepstorff, P., Bennett, E.P., Nielsen, P.A., Hollingsworth, M.A., Burchell, J., Taylor-Papadimitriou, J., and Clausen, H. (1997) Substrate specificities of three members of the human UDP-N-acetyl-alpha-D-galactosamine: polypeptide N-acetylgalactosaminyltransferase family, GalNAc-T1, -T2, and -T3. J. Biol. Chem., 272, 23503–23514.
Wang, C., Eufemi, M., Turano, C., and Giartosio, A. (1996) Influence of the carbohydrate moiety on the stability of glycoproteins. Biochemistry, 35, 7299–7307.[CrossRef][Medline]
Wang, D., Liu, S., Trummer, B.J., Deng, C., and Wang, A. (2002) Carbohydrate microarrays for the recognition of cross-reactive molecular markers of microbes and host cells. Nat. Biotechnol., 20, 275–281.[CrossRef][Web of Science][Medline]
Weinstein, J., de Souza-e-Silva, U., and Paulson, J.C. (1982) Purification of a Gal beta 1 to 4GlcNAc alpha 2 to 6 sialyltransferase and a Gal beta 1 to 3(4)GlcNAc alpha 2 to 3 sialyltransferase to homogeneity from rat liver. J. Biol. Chem., 257, 13835–13844.
Wells, L. and Hart, G.W. (2003) O-GlcNAc turns twenty: functional implications for post-translational modification of nuclear and cytosolic proteins with a sugar. FEBS Lett., 546, 154–158.[CrossRef][Web of Science][Medline]
Yoshida, Y., Kojima, N., Kurosawa, N., Hamamoto, T., and Tsuji, S. (1995) Molecular cloning of Sia alpha 2,3Gal beta 1,4GlcNAc alpha 2,8-sialyltransferase from mouse brain. J. Biol. Chem., 270, 14628–14633.
Young, W.W. Jr., Holcomb, D.R., Ten Hagen, K.G., and Tabak, L.A. (2003) Expression of UDP-GalNAc: polypeptide N-acetylgalactosaminyltransferase isoforms in murine tissues determined by real-time PCR: a new view of a large family. Glycobiology, 13, 549–557.
Zara, J., Hagen, F.K., Ten Hagen, K.G., Van Wuyckhuyse, B.C., and Tabak, L.A. (1996) Cloning and expression of mouse UDP-GalNAc: polypeptide N-acetylgalactosaminyltransferase-T3. Biochem. Biophys. Res. Commun., 228, 38–44.[CrossRef][Web of Science][Medline]
Zhang, J.Q., Biedermann, B., Nitschke, L., and Crocker, P.R. (2004) The murine inhibitory receptor mSiglec-E is expressed broadly on cells of the innate immune system whereas mSiglec-F is restricted to eosinophils. Eur. J. Immunol., 34, 1175–1184.[CrossRef][Web of Science][Medline]
CiteULike 
Connotea 
Del.icio.us What's this?
This article has been cited by other articles:
|
|
 |
|
  M. L. Montpetit, P. J. Stocker, T. A. Schwetz, J. M. Harper, S. A. Norring, L. Schaffer, S. J. North, J. Jang-Lee, T. Gilmartin, S. R. Head, et al. Regulated and aberrant glycosylation modulate cardiac electrical signaling PNAS, September 22, 2009; 106(38): 16517 - 16522. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 F. Guillerme-Bosselut, L. Forestier, C. Jayat-Vignoles, J.-L. Vilotte, I. Popa, J. Portoukalian, A. Le Dur, H. Laude, R. Julien, and P.-F. Gallet Glycosylation-related gene expression profiling in the brain and spleen of scrapie-affected mouse Glycobiology, August 1, 2009; 19(8): 879 - 889. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 A. V. Nairn, W. S. York, K. Harris, E. M. Hall, J. M. Pierce, and K. W. Moremen Regulation of Glycan Structures in Animal Tissues: TRANSCRIPT PROFILING OF GLYCAN-RELATED GENES J. Biol. Chem., June 20, 2008; 283(25): 17298 - 17313. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
Z. S Kawar, T. K Johnson, S. Natunen, J. B Lowe, and R. D Cummings PSGL-1 from the murine leukocytic cell line WEHI-3 is enriched for core 2-based O-glycans with sialyl Lewis x antigen Glycobiology, June 1, 2008; 18(6): 441 - 446. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
H. Yagi, M. Nakagawa, N. Takahashi, S. Kondo, M. Matsubara, and K. Kato Neural complex-specific expression of xylosyl N-glycan in Ciona intestinalis Glycobiology, February 1, 2008; 18(2): 145 - 151. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. S. Powlesland, T. Fisch, M. E. Taylor, D. F. Smith, B. Tissot, A. Dell, S. Pohlmann, and K. Drickamer A Novel Mechanism for LSECtin Binding to Ebola Virus Surface Glycoprotein through Truncated Glycans J. Biol. Chem., January 4, 2008; 283(1): 593 - 602. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 M. Bax, J. J. Garcia-Vallejo, J. Jang-Lee, S. J. North, T. J. Gilmartin, G. Hernandez, P. R. Crocker, H. Leffler, S. R. Head, S. M. Haslam, et al. Dendritic Cell Maturation Results in Pronounced Changes in Glycan Expression Affecting Recognition by Siglecs and Galectins J. Immunol., December 15, 2007; 179(12): 8216 - 8224. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
G. Alvarez-Manilla, N. L. Warren, T. Abney, J. Atwood III, P. Azadi, W. S York, M. Pierce, and R. Orlando Tools for glycomics: relative quantitation of glycans by isotopic permethylation using 13CH3I Glycobiology, July 1, 2007; 17(7): 677 - 687. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
A. Ishii, T. Ikeda, S. Hitoshi, I. Fujimoto, T. Torii, K. Sakuma, S.-i. Nakakita, S. Hase, and K. Ikenaka Developmental changes in the expression of glycogenes and the content of N-glycans in the mouse cerebral cortex Glycobiology, March 1, 2007; 17(3): 261 - 276. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
T. Kudo, T. Fujii, S. Ikegami, K. Inokuchi, Y. Takayama, Y. Ikehara, S. Nishihara, A. Togayachi, S. Takahashi, K. Tachibana, et al. Mice lacking {alpha}1,3-fucosyltransferase IX demonstrate disappearance of Lewis x structure in brain and increased anxiety-like behaviors Glycobiology, January 1, 2007; 17(1): 1 - 9. [Abstract] [Full Text] [PDF]
|
|
|
|
 |
|
 R. R. Drake, E. E. Schwegler, G. Malik, J. Diaz, T. Block, A. Mehta, and O. J. Semmes Lectin Capture Strategies Combined with Mass Spectrometry for the Discovery of Serum Glycoprotein Biomarkers Mol. Cell. Proteomics, October 1, 2006; 5(10): 1957 - 1967. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
E. M. Comelli, M. Sutton-Smith, Q. Yan, M. Amado, M. Panico, T. Gilmartin, T. Whisenant, C. M. Lanigan, S. R. Head, D. Goldberg, et al. Activation of Murine CD4+ and CD8+ T Lymphocytes Leads to Dramatic Remodeling of N-Linked Glycans J. Immunol., August 15, 2006; 177(4): 2431 - 2440. [Abstract] [Full Text] [PDF]
|
|
|
|
|
|
R. Raman, M. Venkataraman, S. Ramakrishnan, W. Lang, S. Raguram, and R. Sasisekharan Advancing glycomics: Implementation strategies at the Consortium for Functional Glycomics Glycobiology, May 1, 2006; 16(5): 82R - 90R. [Abstract] [Full Text] [PDF]
|
|
| ||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||