Recruitment of O-GlcNAc transferase to promoters by corepressor mSin3A: coupling protein O-GlcNAcylation to transcriptional repression

Transcription factors and RNA polymerase II can be modified by O-linked N-acetylglucosamine (O-GlcNAc) monosaccharides at serine or threonine residues, yet the precise functional roles of this modification are largely unknown. Here, we show that O-GlcNAc transferase (OGT), the enzyme that catalyzes this posttranslational modification, interacts with a histone deacetylase complex by binding to the corepressor mSin3A. Functionally, OGT and mSin3A cooperatively repress transcription in parallel with histone deacetylation. We propose that mSin3A targets OGT to promoters to inactivate transcription factors and RNA polymerase II by O-GlcNAc modification, which acts in concert with histone deacetylation to promote gene silencing in an efficient and specific manner.

A novel beta1,3-N-acetylglucosaminyltransferase involved in invasion of cancer cells as assayed in vitro

Using a two-step screening system for genes involved in tissue invasion [Kataoka et al., Cancer Lett. 163(2) (2001) 213], we identified a cDNA whose expression level was higher in mouse placenta at later stages of gestation and in sublines of cancer cells with low degrees of invasiveness. The deduced amino acid sequence showed relatively high similarity with beta1,3-N-acetylglucosaminyltransferase2 approximately 5 (beta3GnT2 approximately 5), and the protein was therefore named beta3GnT7. A possible human ortholog was identified and its chromosomal locus was determined to be 2q37.1. In the mouse, beta3GnT7 was most strongly expressed in the placenta and colon. Moderate amounts of mRNA were detected in the lung, stomach, small intestine, and kidney. The expression of beta3GnT7 was very weak in the cerebrum, cerebellum, heart, and testis. Transfection of the antisense oligonucleotide significantly enhanced the motility of a lung cancer cell line (KLN205-MUC1) in a monolayer compared to the controls. Furthermore, the antisense oligonucleotide increased the number of cells that invaded the matrix-coated membrane in an in vitro invasion model. These results indicate that beta3GnT7 may play a role in preventing cells from migrating out of the original tissues and invading surrounding tissues.

A secreted type of beta 1,6-N-acetylglucosaminyltransferase V (GnT-V) induces tumor angiogenesis without mediation of glycosylation: a novel function of GnT-V distinct from the original glycosyltransferase activity

Angiogenesis is the first regulatory step of tumor progression. Herein, we report on some findings that show that beta1,6-N-acetylglucosaminyltransferase V (GnT-V) functions as an inducer of angiogenesis that has a novel and completely different function from the original function of glycosyltransferase. A secreted type of GnT-V protein itself promoted angiogenesis in vitro and in vivo at physiological concentrations. The highly basic domain of GnT-V induced the release of fibroblast growth factor-2 from heparan sulfate proteoglycan on the cell surface and/or extracellular matrix, leading to angiogenesis. These findings provide some novel information on the relationship between GnT-V and tumor metastasis. The inhibition of GnT-V secretion or its expression represents a novel potential strategy for the inhibition of tumor angiogenesis.

CD45 modulates galectin-1-induced T cell death: regulation by expression of core 2 O-glycans

Galectin-1 induces death of immature thymocytes and activated T cells. Galectin-1 binds to T cell-surface glycoproteins CD45, CD43, and CD7, although the precise roles of each receptor in cell death are unknown. We have determined that CD45 can positively and negatively regulate galectin-1-induced T cell death, depending on the glycosylation status of the cells. CD45(+) BW5147 T cells lacking the core 2 beta-1,6-N-acetylglucosaminyltransferase (C2GnT) were resistant to galectin-1 death. The inhibitory effect of CD45 in C2GnT(-) cells appeared to require the CD45 cytoplasmic domain, because Rev1.1 cells expressing only CD45 transmembrane and extracellular domains were susceptible to galectin-1 death. Moreover, treatment with the phosphotyrosine-phosphatase inhibitor potassium bisperoxo(1,10-phenanthroline)oxovanadate(V) enhanced galectin-1 susceptibility of CD45(+) T cell lines, but had no effect on the death of CD45(-) T cells, indicating that the CD45 inhibitory effect involved the phosphatase domain. Expression of the C2GnT in CD45(+) T cell lines rendered the cells susceptible to galectin-1, while expression of the C2GnT in CD45(-) cells had no effect on galectin-1 susceptibility. When CD45(+) T cells bound to galectin-1 on murine thymic stromal cells, only C2GnT(+) T cells underwent death. On C2GnT(+) cells, CD45 and galectin-1 co-localized in patches on membrane blebs while no segregation of CD45 was seen on C2GnT(-) T cells, suggesting that oligosaccharide-mediated clustering of CD45 facilitated galectin-1-induced cell death.

Differential requirements for core2 glucosaminyltransferase for endothelial L-selectin ligand function in vivo

L-selectin is a calcium-dependent lectin on leukocytes mediating leukocyte rolling in high endothelial venules and inflamed microvessels. Many selectin ligands require modification of glycoproteins by leukocyte core2 beta1,6-N-acetylglucosaminyltransferase (Core2GlcNAcT-I). To test the role of Core2GlcNAcT-I for L-selectin ligand biosynthesis, we investigated leukocyte rolling in venules of untreated and TNF-alpha-treated cremaster muscles and in Peyer's patch high endothelial venules (HEV) of Core2GlcNAcT-I null (core2(-/-)) mice. In the presence of blocking mAbs against P- and E-selectin, L-selectin-mediated leukocyte rolling was almost completely abolished in cremaster muscle venules of core2(-/-) mice, but not littermate control mice. By contrast, leukocyte rolling in Peyer's patch HEV was not significantly different between core2(-/-) and control mice. To probe L-selectin ligands more directly, we injected L-selectin-coated beads. These beads showed no rolling in cremaster muscle venules of core2(-/-) mice, but significant rolling in controls. In Peyer's patch HEV, beads coated with a low concentration of L-selectin showed reduced rolling in core2(-/-) mice. Beads coated with a 10-fold higher concentration of L-selectin rolled equivalently in core2(-/-) and control mice. Our data show that endothelial L-selectin ligands relevant for rolling in inflamed microvessels of the cremaster muscle are completely Core2GlcNAcT-I dependent. In contrast, L-selectin ligands in Peyer's patch HEV are only marginally affected by the absence of Core2GlcNAcT-I, but are sufficiently functional to support L-selectin-dependent leukocyte rolling in Core2GlcNAcT-I-deficient mice.

Cutting edge: differential requirements for Stat4 in expression of glycosyltransferases responsible for selectin ligand formation in Th1 cells

A role for Stat4 in IL-12-induced up-regulation of selectin ligands on Th1 cells was explored. Th1 cells generated from Stat4(-/-) mice exhibited no IL-12-inducible P-selectin ligands, no up-regulation of core 2 beta1,6-glucosaminyltransferase I (C2GlcNAcT-I), and low levels of the Th1 transcription factor T-bet. In contrast, Stat4(-/-) Th1 cells exhibited only a partial defect in expression of IL-12-inducible E-selectin ligands and expressed equivalently high levels of alpha1,3-fucosyltransferase VII (FucT-VII) as wild-type Th1 cells. FucT-VII expression was induced by T cell activation, and was enhanced by IL-12 independently of Stat4, whereas C2GlcNAcT-I up-regulation was mediated exclusively by IL-12, acting through Stat4. These data show that FucT-VII and C2GlcNAcT-I are controlled through distinct pathways and imply the existence of at least one other IL-12-inducible glycosyltransferase required for E-selectin and possibly P-selectin ligand formation in Th1 cells.

Core 2 beta1,6-N-acetylglucosaminyltransferases and alpha1,3-fucosyltransferases regulate the synthesis of O-glycans on selectin ligands on oral cavity carcinoma cells

Selectin-dependent cell binding has importance in the extravasation of blood-circulating tumor cells and in the generation of metastases. Cell surface glycoproteins decorated with sialylated, fucosylated epitopes, such as sialyl Lewis(x) (sLe(x)), are ligands for selectins. Not only terminal sLe(x) moieties but also proximal core structures contribute to the formation of binding epitopes for selectins. Core 2 beta1,6-N-acetylglucosaminyltransferases (C2GnT) and alpha1,3-fucosyltransferases (alpha1,3-FucT) have been suggested to be the rate-limiting enzymes in the synthesis of selectin ligands. We analyzed oral cavity epithelial carcinoma cell lines and showed their expression of RNA transcripts for C2GnT and alpha1,3-FucT, identified alpha1,3-FucT enzyme activities, and analyzed the cell surface sLe(x) expression levels. Neither the pattern of expressed enzymes nor the alpha1,3-FucT activity directly predicted the binding capacity of E-selectin. However, only the sLe(x)-expressing cell lines were capable of binding to E-selectin, but not to P-selectin, thus putatively promoting the selectin-mediated metastasis. These findings suggest that C2GnT in combination with alpha1,3-Fuc-T contribute to the selectin-mediated metastasis in oral cavity carcinomas.

Relationship between metastasis-associated phenotypes and N-glycan structure of surface glycoproteins in human hepatocarcinoma cells

PURPOSE

To study the relation of N-glycan structure on cell surface glycoproteins to the metastatic phenotypes.

METHODS

Two human hepatocarcinoma 7721 cell lines transfected with sense or antisense cDNA of GnT-V, named GnT-V/7721 and GnT-V-AS/7721, respectively, were adopted, because GnT-V can change the antennary number and the content of the beta 1,6 GlcNAc branch in N-glycans. The effects of over- and under-expression of GnT-V on the metastasis-associated phenotype of the transfected cells were investigated and compared with the cells mock-transfected with the plasmid vector.

RESULTS

In GnT-V/7721 cells, GnT-V activity was increased by 92% compared with the mock cells. HRP-labeled lectin staining of transfected cells showed elevated intensity with HRP-L-PHA and reduced intensity with HRP-ConA, suggesting the increased antennary number and content of the beta 1,6 GlcNAc branch in N-glycans. Analysis of the N-glycan structure of [3H]-labeled glycopeptides prepared from cell-surface [3H] glycoproteins using DSA-affinity chromatography also revealed the above change of the N-glycan structure in a more quantitative manner. GnT-V/7721 cells showed a suppressed cell attachment to fibronectin (Fn) or laminin (Ln), and increased cell migration and invasion through matrigel. In contrast, GnT-V-AS/7721 cells showed reduction of both GnT-V activity and content of the beta 1,6 branch in N-linked glycans, elevation of cell attachment to Fn or Ln, and decline of cell migration and invasion through matrigel. These changes were just the opposite to those in GnT-V/7721 cells.

CONCLUSIONS

The alteration of N-glycan structure in surface glycoproteins resulting from the activity change of GnT-V contributes to the alterations in metastasis-associated phenotypes. The product of GnT-V, the beta 1,6 GlcNAc branch in N-linked glycans, is a structural factor of adhesion inhibition and invasion promotion. GnT-V is, therefore, closely related to cancer metastasis and its over-expression is an important molecular mechanism of metastasis.

rib-2, a Caenorhabditis elegans homolog of the human tumor suppressor EXT genes encodes a novel alpha1,4-N-acetylglucosaminyltransferase involved in the biosynthetic initiation and elongation of heparan sulfate

The proteins encoded by the EXT1, EXT2, and EXTL2 genes, members of the hereditary multiple exostoses gene family of tumor suppressors, are glycosyltransferases required for the heparan sulfate biosynthesis. Only two homologous genes, rib-1 and rib-2, of the mammalian EXT genes were identified in the Caenorhabditis elegans genome. Although heparan sulfate is found in C. elegans, the involvement of the rib-1 and rib-2 proteins in heparan sulfate biosynthesis remains unclear. In the present study, the substrate specificity of a soluble recombinant form of the rib-2 protein was determined and compared with those of the recombinant forms of the mammalian EXT1, EXT2, and EXTL2 proteins. The present findings revealed that the rib-2 protein was a unique alpha1,4-N-acetylglucosaminyltransferase involved in the biosynthetic initiation and elongation of heparan sulfate. In contrast, the findings confirmed the previous observations that both the EXT1 and EXT2 proteins were heparan sulfate copolymerases with both alpha1,4-N-acetylglucosaminyltransferase and beta1,4-glucuronyltransferase activities, which are involved only in the elongation step of the heparan sulfate chain, and that the EXTL2 protein was an alpha1,4-N-acetylglucosaminyltransferase involved only in the initiation of heparan sulfate synthesis. These findings suggest that the biosynthetic mechanism of heparan sulfate in C. elegans is distinct from that reported for the mammalian system.

Metastases: the glycan connection

An association between protein glycosylation and tumorigenesis has been recognized for over 10 years. Associations linking the importance of glycosylation events to tumor biology, especially the progression to metastatic disease, have been noted over many years, Recently, a mouse model in which beta1,6-N-acetylglucosaminyltransferase V (a rate-limiting enzyme in the N-glycan pathway) has been knocked out, was used to demonstrate the importance of glycosylation in tumor progression. By crossing mice lacking this enzyme with a transgenic mouse model of metastatic breast cancer, metastatic progression of the disease was dramatically reduced. These experiments provide in vivo evidence for the role of N-linked glycosylation in metastatic breast cancer and have significant implications for the development of new treatment strategies.

[Multiple exostoses]

Multiple exostoses is a hereditary disease characterized by multiple osteocartilagenous tumors, of which the histological structures are similar to those of normal epiphyses. Genetic linkage has identified three different loci for this disease: EXT1 on 8q, EXT2 on 11p, and EXT3 on 19p. The EXT1 and EXT2 genes were recently isolated and mutation analyses have been performed in a number of patients with different ethnic backgrounds. The data indicate that mutations of these genes occurred in broad regions of each gene, and the loss-of-function mutations were predominant, although there were some missense mutations that may create functionally defective protein. Tumor cells were shown to be homozygous for the mutant allele, which is consistent with the concept of these genes as tumor suppressor genes. Recent progress for the functional analyses has disclosed that these genes encode the protein with glycosyltransferase activity and regulate the diffusion of Hedgehog protein, which is the key molecule for the skeletal development. Further analyses of these genes may provide us with the knowledge for the development of epiphyses, and may open the new research field for the regeneration of epiphyses.

New evidence for an extra-hepatic role of N-acetylglucosaminyltransferase III in the progression of diethylnitrosamine-induced liver tumors in mice

N-acetylglucosaminyltransferase III (GlcNAc-TIII) is encoded by the Mgat3 gene and catalyzes the addition of the bisecting GlcNAc to the core of N-glycans. Mice lacking GlcNAc-TIII due to the insertion mutation Mgat3tmlPst (termed Mgat3neo), exhibit retarded progression of liver tumors induced by diethylnitrosamine (DEN; M. Bhaumik et al, Cancer Res., 58: 2881-2887, 1998). This phenotype seemed to be due to a reduction, in activity or amount, of a circulating glycoprotein(s) that enhances DEN-induced liver tumor progression. Here, we provide new evidence to support this hypothesis. First, we show that mice with a deletion mutation of the Mgat3 gene coding exon (Mgat3tmlJxm, termed Mgat3delta) also exhibit retarded progression of DEN-induced liver tumors. At 7 months there was a significant decrease in liver weight (approximately 27%; P < 0.01), reflecting reduced tumor burden in Mgat3delta/delta mice. In addition, tumors were generally fewer and smaller, and histological changes were less severe in Mgat3delta/delta livers. Therefore, tumor progression is retarded in mice with two different null mutations in the Mgat3 gene. Second, we show that the development of DEN-induced tumors is unaltered by high levels of GlcNAc-TIII in the liver of transgenic mice. The Mgat3 gene coding exon under the control of the major urinary protein (MUP) promoter was used to generate transgenic mice that express GlcNAc-TIII in liver. Following DEN injection and phenobarbitol treatment, however, no significant differences were observed between MUP/Mgat3 transgenic and control mice in either tumor numbers or liver weight. The combined data provide strong evidence that retarded progression of tumors in mice lacking GlcNAc-TIII is due to the absence of the bisecting GlcNAc residue on N-glycans of a circulating glycoprotein(s) from a tissue other than liver.

Overexpression of core 2 N-acetylglycosaminyltransferase enhances cytokine actions and induces hypertrophic myocardium in transgenic mice

Elevated levels of glycocojugates, commonly observed in the myocardium of diabetic animals and patients, are postulated to contribute to the myocardial dysfunction in diabetes. Previously, we reported that UDP-GlcNAc: Galbeta1-3GalNAcalphaRbeta1-6-N-acetylglucosaminyltransferas e (core 2 GlcNAc-T), a developmentally regulated enzyme of O-linked glycans biosynthesis pathway, is specifically increased in the heart of diabetic animals and is regulated by hyperglycemia and insulin. In this study, transgenic mice overexpressing core 2 GlcNAc-T with severe increase in cardiac core 2 GlcNAc-T activities were normal at birth but showed progressive and significant cardiac hypertrophy at 6 months of age. The heart of transgenic mice showed elevation of sialylated O-glycan and increases of c-fos gene expression and AP-1 activity, which are characteristics of cardiac stress. Furthermore, transfection of PC12 cells with core 2 GlcNAc-T also induced c-fos promoter activation, mitogen activated-protein kinase (MAPK) phosphorylation, Trk receptor glycosylation, and cell differentiation. These results suggested a novel role for core 2 GlcNAc-T in the development of diabetic cardiomyopathy and modulation of the MAP kinase pathway in the heart.-Koya, D., Dennis, J. W., Warren, C. E., Takahara, N., Schoen, F. J., Nishio, Y., Nakajima, T., Lipes, M. A., King, G. L. Overexpression of core 2 N-acetylglycosaminyltransferase enhances cytokine actions and induces hypertrophic myocardium in transgenic mice.

Progression of hepatic neoplasms is severely retarded in mice lacking the bisecting N-acetylglucosamine on N-glycans: evidence for a glycoprotein factor that facilitates hepatic tumor progression

The glycosyltransferase termed GlcNAc-TIII is dedicated to the transfer of a single N-acetylglucosamine (GlcNAc) residue (the bisecting GlcNAc), to a subset of N-glycans in glycoproteins. The addition of this GlcNAc is differentially regulated during development and is induced in certain cancers, particularly in hepatic tumorigenesis. To investigate a functional role for the bisecting GlcNAc in the development of liver cancer, the Mgat3 gene that codes for GlcNAc-TIII, was inactivated by targeted gene disruption, and the susceptibility of Mgat3-/- mice to tumor induction was tested. After a single injection with diethylnitrosamine and subsequent treatment with phenobarbitol for 6 months, Mgat3+/+ and Mgat3+/- mice had grossly enlarged livers that contained numerous tumors. By stark contrast, Mgat3-/- mice had livers of normal size, and only 50% of mice had one to four small tumors. However, histological examination showed that Mgat3-/- livers had significant numbers of basophilic foci, and by 10-12 months after diethylnitrosamine injection, tumors had developed in Mgat3-/- mice. Therefore, initiation occurred in Mgat3-/- mice but progression was severely retarded. Assays for Mgat3 gene expression in tumor tissue gave an unexpected result. In contrast to the situation in the rat, hepatic tumor formation in the mouse was not accompanied by a dramatic increase of GlcNAc-TIII activity nor of glycoproteins with a bisecting GlcNAc, nor of Mgat3 gene expression in tumor tissue from wild-type mice. The data suggest that a glycoprotein factor with the bisecting GlcNAc facilitates tumor progression in liver. In the absence of the bisecting GlcNAc in Mgat3-/- mice, the factor is reduced in activity, and tumor progression is severely retarded.

Rhizobium nodulation protein NodC is an important determinant of chitin oligosaccharide chain length in Nod factor biosynthesis

Synthesis of chitin oligosaccharides by NodC is the first committed step in the biosynthesis of rhizobial lipochitin oligosaccharides (LCOs). The distribution of oligosaccharide chain lengths in LCOs differs between various Rhizobium species. We expressed the cloned nodC genes of Rhizobium meliloti, R. leguminosarum bv. viciae, and R. loti in Escherichia coli. The in vivo activities of the various NodC proteins differed with respect to the length of the major chitin oligosaccharide produced. The clearest difference was observed between strains with R. meliloti and R. loti NodC, producing chitintetraose and chitinpentaose, respectively. In vitro experiments, using UDP-[14C]GlcNAc as a precursor, show that this difference reflects intrinsic properties of these NodC proteins and that it is not influenced by the UDP-GlcNAc concentration. Analysis of oligosaccharide chain lengths in LCOs produced by a R. leguminosarum bv. viciae nodC mutant, expressing the three cloned nodC genes mentioned above, shows that the difference in oligosaccharide chain length in LCOs of R. meliloti and R. leguminosarum bv. viciae is due only to nodC. The exclusive production of LCOs which contain a chitinpentaose backbone by R. loti strains is not due to NodC but to end product selection by Nod proteins involved in further modification of the chitin oligosaccharide. These results indicate that nodC contributes to the host specificity of R. meliloti, a conclusion consistent with the results of several studies which have shown that the lengths of the oligosaccharide backbones of LCOs can strongly influence their activities on host plants.

Organization of the human beta-1,2-N-acetylglucosaminyltransferase I gene (MGAT1), which controls complex and hybrid N-glycan synthesis

UDP-GlcNAc: alpha-3-D-mannoside beta-1,2-N-acetylglucosaminyltransferase I (EC 2.4.1.101; GlcNAc-T I) is a medial-Golgi enzyme which catalyses the first step in the conversion of oligomannose-type to N-acetyl-lactosamine- and hybrid-type N-glycans and is essential for normal embryogenesis in the mouse. Previous work indicated the presence of at least two exons in the human GlcNAc-T I gene MGAT1, exon 2 containing part of the 5' untranslated region and the complete coding and 3' untranslated regions, and exon 1 with the remainder of the 5' untranslated region. We now report the cloning and sequencing of a human genomic DNA fragment containing exon 1, which is between 5.6 and 15 kb upstream of exon 2. Transient transfection, ribonuclease protection and reverse transcriptase-mediated PCR indicated the absence of transcription start sites in intron 1 between exons 1 and 2. Northern analysis, ribonuclease protection, primer extension analysis and rapid amplification of 5'-cDNA ends showed that there are multiple transcription start sites for exon 1 compatible with the expression by several human cell lines and tissues of two transcripts, a broad band ranging in size from 2.7 to 3.0 kb and a sharper band at 3.1 kb. The 5' flanking region of exon 1 has a GC content of 81% and has no canonical TATA or CCAAT boxes but contains potential binding sites for transcription factors Sp1, GC-binding factor and epidermal growth factor receptor-specific transcription factor. Chloramphenicol acetyltransferase (CAT) expression was observed on transient transfection into HeLa cells of a fusion construct containing the gene for CAT and a genomic DNA fragment from the 5' flanking region of exon 1. It is concluded that MGAT1 is a typical housekeeping gene although there is, in addition, tissue-specific expression of the larger 3.1 kb transcript.

Core2 beta-1,6-N-acetylglucosaminyltransferase enzyme activity is critical for P-selectin glycoprotein ligand-1 binding to P-selectin

P-selectin glycoprotein ligand-1 (PSGL-1) is a high-affinity counterreceptor for P-selectin on myeloid cells and activated T-cells. In addition, PSGL-1 can serve, both in vitro and in vivo, as an E-selectin ligand. Appropriate glycosylation of PSGL-1 is crucial for binding to P-selectin. Functional PSGL-1 is known to bear sialyl lewis X (SLex) or a closely related oligosaccharide. In this study, we show that Chinese hamster ovary (CHO) cells expressing PSGL-1 and fucosyltransferase show a dramatic increase in binding to P-selectin when transfected with "core2" transferase, the enzyme that initiates branching of O-linked glycans. Moreover, only PSGL-1 from core2 transfectant CHO cells can be affinity-captured with P-selectin, suggesting that branched O-linked glycans are required for high-affinity binding to P-selectin. Analysis of PSGL-1-derived O-linked oligosaccharides produced in core2 transfected cells shows the presence of more elaborated glycans. Interestingly, transfection of core2 in these cells does not alter binding to E-selectin.

Essential role for complex N-glycans in forming an organized layer of bronchial epithelium

Mice lacking the complex subset of N-glycans due to inactivation of the Mgat1 gene die at mid-gestation, making it difficult to identify specific biological functions for this class of cell surface carbohydrates. To circumvent this embryonic lethality and to uncover tissue-specific functions for complex N-glycans, WW6 embryonic stem cells with inactivated Mgat1 alleles were tracked in chimeric embryos. The Mgat1 gene encodes N-acetylglucosaminyltransferase I (Glc-NAc-TI; EC 2.4.1.101), the transferase that initiates the synthesis of complex N-glycans. WW6 cells carry an inert beta-globin transgene that allows their identification in chimeras by DNA-DNA in situ hybridization. Independent Mgat1-/- and Mgat1+/- mutant WW6 isolates contributed like parent WW6 cells to the tissues of embryonic day (E) 10.5 to E16.5 chimeras. However, a cell type-specific difference was observed in lung. Homozygous null Mgat1-/- WW6 cells did not contribute to the epithelial layer in more than 99% bronchi. This deficiency was corrected by transfection of a Mgat1 transgene. Interestingly, heterozygous Mgat1+/- WW6 cells were also deficient in populating the layer of bronchial epithelium. Furthermore, examination of lung bud in E9.5 Mgat1-/- mutant embryos showed complete absence of an organized epithelial cell layer in the bronchus. Thus, complex N-glycans are required to form a morphologically recognizable bronchial epithelium, revealing an in vivo, cell type-specific function for this class of N-glycans.