Trafficking and localization studies of recombinant alpha1, 3- fucosyltransferase VI stably expressed in CHO cells

Peripheral alpha1,3-fucosylation of glycans occurs by the action of either one of five different alpha1,3-fucosyltransferases (Fuc-Ts) cloned to date. Fuc-TVI is one of the alpha1,3-fucosyltransferases which is capable to synthesize selectin ligands. The major alpha1, 3- fucosyltransferase activity in human plasma is encoded by the gene for fucosyltransferase VI, which presumably originates from liver cells. While the sequence, chromosomal localization, and kinetic properties of Fuc-TVI are known, immunocytochemical localization and trafficking studies have been impossible because of the lack of specific antibodies. Here we report on the development and characterization of a peptide-specific polyclonal antiserum monospecific to Fuc-TVI and an antiserum to purified soluble recombinant Fuc-TVI crossreactive with Fuc-TIII and Fuc-TV. Both antisera were applied for immunodetection in stably transfected CHO cells expressing the full-length form of this enzyme (CHO clone 61/11). Fuc-TVI was found to be a resident protein of the Golgi apparatus. In addition, more than 30% of cell-associated and released enzyme activity was found in the medium. Maturation and release of Fuc-TVI was analyzed in metabolically labeled CHO 61/11 cells followed by immunoprecipitation. Fuc-TVI occurred in two forms of 47 kDa and 43 kDa bands, while the secreted form was detected as a 43 kDa. These two different intracellular forms arose by posttranslational modification, as shown by pulse-chase experiments. Fuc-TVI was released to the supernatant by proteolytic cleavage as a partially endo-H resistant glycoform.      

Immunocytochemical localization of alpha2,3(N)-sialyltransferase (ST3Gal III) in cell lines and rat kidney tissue sections: evidence for golgi and post-golgi localization

Sialylation is a biosynthetic process occurring in the trans compartments of the Golgi apparatus. Corresponding evidence is based on localization and biochemical studies of alpha2, 6(N)-sialyltransferase (ST6Gal I) as previously reported. Here we describe generation and characterization of polyclonal antibodies to recombinant rat alpha2,3(N)-sialyltransferase (ST3Gal III) expressed as a soluble enzyme in Sf9 cells or as a beta-galactosidase-human-ST3Gal III fusion- protein from E.coli , respectively. These antibodies were used to localize ST3Gal III by immunofluorescence in various cell lines and rat kidney tissue sections. In transiently transfected COS cells the antibodies directed to soluble sialyltransferase or the sialyltransferase portion of the fusion-protein only recognized the recombinant antigen retained in the endoplasmic reticulum. However, an antibody fraction crossreactive with beta-galactosidase recognized natively expressed ST3Gal III which was found to be colocalized with beta1, 4-galactosyltransferase in the Golgi apparatus of several cultured cell lines. Antibodies affinity purified on the beta- galactosidase-ST3Gal III fusion-protein column derived from both antisera have then been used to localize the enzyme in perfusion-fixed rat kidney sections. We found strong staining of the Golgi apparatus of tubular epithelia and a brush-border-associated staining which colocalized with cytochemical staining of the H+ATPase. This subcellular localization was not observed for ST6Gal I which localized to the Golgi apparatus. These data show colocalization in the Golgi apparatus and different post-Golgi distributions of the two sialyltransferases.      

Colocalization of β1,4galactosyltransferase with mannose 6-phosphate receptor in monensin-induced TGN-derived structures

Previously, we demonstrated that beta 1,4galactosyltransferase (gal-T1) reversibly segregates from alpha 2,6sialyltransferase (ST6Gal) to swollen vesicles after monensin treatment of the cells. To further explore this phenomenon, we investigated the response to monensin of various Golgi proteins. Within 30 min of monensin treatment, gal-T1 moved from the Golgi apparatus, as defined by localization of giantin, to swollen vesicles whereas ST6Gal, alpha 2,3(N)sialyltransferase, mannosidase II, and N-acetylgalactosaminyltransferase 2 remained associated with the Golgi apparatus. Stably transfected CHO cells exhibited a similar phenomenon of monensin-induced displacement of recombinant gal-T1 to swollen vesicles while recombinant ST6Gal remained colocalized with endogenously expressed giantin. Gal-T1 and the cation-insensitive mannose 6-phosphate receptor colocalized in swollen vesicles as observed at both light and electron microscopic levels. When monensin was replaced by chloroquine, gal-T1 remained arrested in swollen vesicles. Brefeldin A treatment known to cause relocation of Golgi-associated gal-T1 to the endoplasmic reticulum had no effect on gal-T1 trapped in swollen vesicles. This evidence suggests that monensin blocks gal-T1 trafficking in post-Golgi structures and argues against swelling of gal-T1-containing trans Golgi cisternae as previously assumed.     

The acceptor substrate specificity of human beta4-galactosyltransferase V indicates its potential function in O-glycosylation.

In order to assess the function of the different human UDP-Gal:GlcNAc beta4-galactosyltransferases, the cDNAs of two of them, beta4-GalT I and beta4-GalT V, were expressed in the baculovirus/insect cell expression system. The soluble recombinant enzymes produced were purified from the medium and used to determine their in vitro substrate specificities. The specific activity of the recombinant beta4-GalT V was more than 15 times lower than that of beta4-GalT I, using GlcNAc beta-S-pNP as an acceptor. Whereas beta4-GalT I efficiently acts on all substrates having a terminal beta-linked GlcNAc, beta4-GalT V appeared to be far more restricted in acceptor usage. Beta4-GalT V acts with high preference on acceptors that contain the GlcNAc beta1-->6GalNAc structural element, as found in O-linked core 2-, 4- and 6-based glycans, but not on substrates related to V-linked or blood group I-active oligosaccharides. These results suggest that beta4-GalT V may function in the synthesis of lacNAc units on O-linked chains, particularly in tissues which do not express beta4-GalT I, such as brain.     

Identification of a high affinity binding protein for the regulatory subunit RII beta of cAMP-dependent protein kinase in Golgi enriched membranes of human lymphoblasts.

Immunocytochemical evidence of an association between the regulatory subunit RII of the cAMP-dependent protein kinase (cAMP-PK) and the Golgi apparatus in several cell types has been reported. In order to identify endogenous Golgi proteins binding RII, a fraction enriched in Golgi vesicles was isolated from human lymphoblasts. Only the RII beta isoform was detected in the Golgi-rich fraction, although RII alpha has also been found to be present in these cells. A 85 kDa RII-binding protein was identified in Golgi vesicles using a [32P]RII overlay of Western blots. The existence of an endogenous RII beta-p85 complex in isolated Golgi vesicles was demonstrated by two independent means: (i) co-immunoprecipitation of both proteins under non-denaturing conditions with an antibody against RII beta and (ii) co-purification of RII beta-p85 complexes on a cAMP-analogue affinity column. p85 was phosphorylated by both endogenous and purified catalytic subunits of cAMP-pKII. Extraction experiments and protease protection experiments indicated that p85 is an integral membrane protein although it partitioned atypically during Triton X-114 phase separation. We propose that p85 anchors RII beta to the Golgi apparatus of human lymphoblasts and thereby defines the Golgi substrate targets most accessible to phosphorylation by C subunit. This mechanism may be relevant to the regulation of processes involving the Golgi apparatus itself, such as membrane traffic and secretion, but also relevant to nearby nuclear events dependent on C subunit.     

An isoform of the Golgi t-SNARE, syntaxin 5, with an endoplasmic reticulum retrieval signal.

The early Golgi t-SNARE (target-membrane-associated soluble-N-ethylmaleimide-sensitive factor attachment protein receptor) syntaxin 5 is thought to specify the docking site for both COPI and COPII coated vesicles originating from the endoplasmic reticulum (ER) and COPI vesicles on the retrograde pathway. We now show that there are two forms of syntaxin 5 that appear to be generated from the same mRNA by alternative initiation of translation. The short form (35 kDa) corresponds to the published sequence. The long form (42 kDa) has an N-terminal cytoplasmic extension containing a predicted type II ER retrieval signal. When grafted onto a reporter molecule, this signal localized the construct to the ER. Biochemical fractionation and immunofluorescence microscopy showed that there was less of the long form in the Golgi apparatus and more in peripheral punctate structures, some of which colocalized with markers of the intermediate compartment. The predicted absence of the long form in budding yeast points to a function unique to higher organisms.     

A vesicular intermediate in the transport of hepatoma secretory proteins from the rough endoplasmic reticulum to the Golgi complex

We have identified a vesicle fraction that contains alpha 1-antitrypsin and other human HepG2 hepatoma secretory proteins en route from the rough endoplasmic reticulum (RER) to the cis face of the Golgi complex. [35S]Methionine pulse-labeled cells were chased for various periods of time, and then a postnuclear supernatant fraction was resolved on a shallow sucrose-D2O gradient. This intermediate fraction has a density lighter than RER or Golgi vesicles. Most alpha 1-antitrypsin in this fraction (P1) bears N-linked oligosaccharides of composition similar to that of alpha 1-antitrypsin within the RER; mainly Man8GlcNac2 with lesser amounts of Man7GlcNac2 and Man9GlcNac2; this suggests that the protein has not yet reacted with alpha-mannosidase-I on the cis face of the Golgi complex. This light vesicle species is the first post-ER fraction to be filled by labeled alpha 1-antitrypsin after a short chase, and newly made secretory proteins enter this compartment in proportion to their rate of exit from the RER and their rate of secretion from the cells: alpha 1-antitrypsin and albumin faster than preC3 and alpha 1-antichymotrypsin, faster, in turn, then transferrin. Deoxynojirimycin, a drug that blocks removal of glucose residues from alpha 1-antitrypsin in the RER and blocks its intracellular maturation, also blocks its appearance in this intermediate compartment. Upon further chase of the cells, we detect sequential maturation of alpha 1- antitrypsin to two other intracellular forms: first, P2, a form that has the same gel mobility as P1 but that bears an endoglycosidase H- resistant oligosaccharide and is found in a compartment--probably the medial Golgi complex--of density higher than that of the intermediate that contains P1; and second, the mature sialylated form of alpha 1- antitrypsin.     

GDP-fucose uptake into the Golgi apparatus during xyloglucan biosynthesis requires the activity of a transporter-like protein other than the UDP-glucose transporter

The molecular mechanisms regulating hemicelluloses and pectin biosynthesis are poorly understood. An important question in this regard is how glycosyltransferases are oriented in the Golgi cisternae, and how nucleotide sugars are made available for the synthesis of the polymers. Here we show that the branching enzyme xyloglucan alpha,1-2 fucosyltransferase (XG-FucTase) from growing pea (Pisum sativum) epicotyls was latent and protected against proteolytic inactivation on intact, right-side-in pea stem Golgi vesicles. Moreover, much of the XG-FucTase activity was membrane associated. These data indicate that XG-FucTase is a membrane-bound luminal enzyme. GDP-Fuc uptake studies demonstrated that GDP-Fuc was taken up into Golgi vesicles in a protein-mediated process, and that this uptake was not competed by UDP-Glc, suggesting that a specific GDP-Fuc transporter is involved in xyloglucan biosynthesis. Once in the lumen, Fuc was transferred onto endogenous acceptors, including xyloglucan. GDPase activity was detected in the lumen of the vesicles, suggesting than the GDP produced upon transfer of Fuc was hydrolyzed to GMP and inorganic phosphate. We suggest than the GDP-Fuc transporter and GDPase may be regulators of xyloglucan fucosylation in the Golgi apparatus from pea epicotyls.     

Molecular cloning of a fourth member of a human alpha (1,3)fucosyltransferase gene family. Multiple homologous sequences that determine expression of the Lewis x, sialyl Lewis x, and difucosyl sialyl Lewis x epitopes.

We and others have previously described the isolation of three human alpha (1,3)fucosyltransferase genes which form the basis of a nascent glycosyltransferase gene family. We now report the molecular cloning and expression of a fourth homologous human alpha (1,3)fucosyltransferase gene. When transfected into mammalian cells, this fucosyltransferase gene is capable of directing expression of the Lewis x (Gal beta 1-->4[Fuc alpha 1-->3]GlcNAc), sialyl Lewis x (NeuNAc alpha 2-->3Gal beta 1-->4 [Fuc alpha 1-->3]GlcNAc), and difucosyl sialyl Lewis x (NeuNAc alpha 2-->3Gal beta 1-->4[Fuc alpha 1-->3]GlcNAc beta 1-->3 Gal beta 1-->4[Fuc alpha 1-->3]GlcNAc) epitopes. The enzyme shares 85% amino acid sequence identity with Fuc-TIII and 89% identity with Fuc-TV but differs substantially in its acceptor substrate requirements. Polymerase chain reaction analyses demonstrate that the gene is syntenic to Fuc-TIII and Fuc-TV on chromosome 19. Southern blot analyses of human genomic DNA demonstrate that these four alpha (1,3)fucosyltransferase genes account for all DNA sequences that cross-hybridize at low stringency with the Fuc-TIII catalytic domain. Using similar methods, a catalytic domain probe from Fuc-TIV identifies a new class of DNA fragments which do not cross-hybridize with the chromosome 19 fucosyltransferase probes. These results extend the molecular definition of a family of human alpha (1,3)fucosyltransferase genes and provide tools for examining fucosyltransferase gene expression.     

Physical membrane displacement: Reconstitution in a cell-free system and relationship to cell growth+

Summary Physical membrane displacement is a process common to all forms of vesicle budding as well as cell enlargement and pleomorphic shape changes. Cell-free reconstitution of membrane budding has been achieved with transitional endoplasmic reticulum fractions from both plants and animals where 50 to 70 nm transition vesicles have been observed to bud from the part-rough, part-smooth membrane elements that define transitional endoplasmic reticulum. This budding phenomenon requires ATP, is facilitated by cytosol and guanine nucleotides, and is both time- and temperature-dependent. The transitional endoplasmic reticulum buds that form when concentrated by preparative free-flow electrophoresis will attach specifically to cis Golgi apparatus membranes immobilized on nitrocellulose as an acceptor compartment. Golgi apparatus membranes derived from the trans compartment do not serve as an efficient acceptor compartment. Transfer of the vesicles once formed is rapid, nearly complete and no longer dependent upon added ATP. Transfer shows a strict temperature dependency corresponding to that of the intact cell where at temperatures of 16°C or below, vesicles form but do not attach to cis Golgi whereas at temperatures of greater than 16°C, vesicles both form and fuse. The principle ATPase of transitional endoplasmic reticulum which may be involved in the budding process has been identified, characterized and isolated. A 38 kDa cis Golgi apparatus associated protein also has been identified as a potential candidate as a docking protein. Transfer between trans Golgi apparatus and the plasma membrane also has been studied by cell-free analysis. Here, transfer has been found to be stimulated by NADH or NADH plus ascorbate. The role of NADH is unknown but the ability of plant and Golgi apparatus to oxidize NADH is inhibited by brefeldin A, a compound known to block membrane trafficking even at the level of the trans Golgi network. NADH oxidase activity of plasma membranes also has been described and is inhibited as well by brefeldin. Recent observations suggest that brefeldin A may block both the formation of vesicles at the trans Golgi apparatus as well as auxin hormone-stimulated cell elongation in plants. This once again raises the possibility of whether or not plant cell elongation is obligatorily mediated by membrane input from the Golgi apparatus. The latter seems unlikely based on two additional lines of evidence. The first is that auxin-induced cell elongation in plants shows no sharp temperature transition over the range of 4 to 24°C, whereas production of secretory vesicles from the trans Golgi apparatus appears to be largely prevented at temperatures of 18°C or less. Secondly, the sodium selective ionophore, monensin, which effectively blocks the formation of functional secretory vesicles at the trans Golgi apparatus, is also largely without effect on auxin-induced cell elongation for periods of 4 h or longer. Taken together the findings suggest that the action of brefeldin A on vesicle budding at the Golgi apparatus and cell enlargement, are not directly correlated but may represent a common action of the drug on some constituent essential to membrane displacement mechanisms.Abbreviations BFA brefeldin A - IAA indole-3-acetic acid; 2, 4-D 2, 4-dichlorophenoxyacetic acid - NSF N-ethylmaleimide-sensitive factor Much of the information summarized in this report was presented as a plenary lecture at the XV International Botanical Congress Tokyo, Yokohama, Japan, August 28–September 3, 1993.     

Evidence for both prelysosomal and lysosomal intermediates in endocytic pathways

Horseradish peroxidase (HRP), an enzyme internalized by fluid phase pinocytosis, has been used to study the process by which pinosome contents are delivered to lysosomes in Chinese hamster ovary cells. Pinosome contents were labeled by allowing cells to internalize HRP for 3-5 min. Following various chase times, cells were either processed for HRP and acid phosphatase (AcPase) cytochemistry or homogenized and fractionated in Percoll gradients. In Percoll gradients, pinosomes labeled by a 3-5 min HRP pulse behaved as a vesicle population more dense than plasma membrane and less dense than lysosomes. In pulse- chase experiments, internalized HRP was chased rapidly (3-6 min chase) to a density position intermediate between the "initial" pinocytic vesicle population and lysosomes. With longer chase periods, a progressive accumulation of HRP in more dense vesicles was observed. Correspondence between the HRP distribution and lysosomal marker distribution was reached after a approximately 1-h chase. By electron microscope cytochemistry of intact cells, the predominant class of HRP- positive vesicles after pulse uptakes or a 3-min chase period was characterized by a peripheral rim of reaction product and was AcPase negative. After 10-120-min chase periods, the predominant class of HRP- positive vesicles was characterized by luminal deposits and HRP activity was frequently observed in multivesicular bodies. HRP-positive vesicles after a 10- or 30-min chase were AcPase-positive. No HRP activity was detected in Golgi apparatus. Together these observations indicate that progressive processing of vesicular components of the vacuolar apparatus occurs at both a prelysosomal and lysosomal stage.     

Colocalisation of the protein tyrosine phosphatases PTP-SL and PTPBR7 with beta4-adaptin in neuronal cells

The mouse gene Ptprr encodes the neuronal protein tyrosine phosphatases PTP-SL and PTPBR7. These proteins differ in their N-terminal domains, with PTP-SL being a cytosolic, membrane-associated phosphatase and PTPBR7 a type I transmembrane protein. In this study, we further explored the nature of the PTP-SL-associated vesicles in neuronal cells using a panel of organelle markers and noted a comparable subcellular distribution for PTP-SL and the beta4-adaptin subunit of the AP4 complex. PTP-SL, PTPBR7 and beta4-adaptin are localised at the Golgi apparatus and at vesicles throughout the cytoplasm. Immunohistochemical analysis demonstrated that PTP-SL, PTPBR7 and beta4-adaptin are all endogenously expressed in brain. Interestingly, coexpression of PTP-SL and beta4-adaptin leads to an altered subcellular localisation for PTP-SL. Instead of the Golgi and vesicle-type staining pattern, still observable for beta4-adaptin, PTP-SL is now distributed throughout the cytoplasm. Although beta4-adaptin was found to interact with the phosphatase domain of PTP-SL and PTPBR7 in the yeast two-hybrid system, it failed to do so in transfected neuronal cells. Our data suggest that the tyrosine phosphatases PTP-SL and PTPBR7 may be involved in the formation and transport of AP4-coated vesicles or in the dephosphorylation of their transmembrane cargo molecules at or near the Golgi apparatus.     

Biosynthesis of the sialyl-Lex determinant carried by type 2 chain glycosphingolipids (IV3NeuAcIII3FucnLc4, VI3NeuAcV3FucnLc6, and VI3NeuAcIII3V3Fuc2nLc6) in human lung carcinoma PC9 cells

The pathway for synthesis of three glycosphingolipids bearing a common sialyl-Lex determinant (NeuAc alpha 2----3Gal beta 1----4[Fuc alpha 1----3]GlcNac beta 1----R) from their type 2 lactoseries precursors has been studied using the 0.2% Triton X-100-soluble fraction from human lung carcinoma PC9 cells. Two enzymes were found to be required for their synthesis: (i) an alpha 1----3 fucosyltransferase, the properties of which have been characterized as being similar to the enzyme from human small cell lung carcinoma NCI-H69 cells (Holmes, E. H., Ostrander, G. K., and Hakomori, S. (1985) J. Biol. Chem. 260, 7619-7627); and (ii) an alpha 2----3 sialyltransferase that was efficiently solubilized by 0.2% Triton X-100 and required divalent metal ions and 0.3% Triton CF-54 for optimal activity at pH 5.9 in cacodylate buffer. Biosynthesis of the sialyl-Lex determinant was shown to proceed via sialylation of nLc6 and nLc4, followed by alpha 1----3 fucosylation at the penultimate GlcNAc residues, based on the following: (i) transfer of NeuAc by PC9 cell sialyltransferase was found only when the nonfucosylated acceptors nLc4 and nLc6 were added, and none of the glycolipids with Lex structure (III3FucnLc4; V3FucnLc6; III3V3Fuc2nLc6) were sialylated; and (ii) the PC9 cell fucosyltransferase was active with both neutral and ganglioside neolacto (type 2 chain) acceptors. Transfer of fucose to VI3NeuAcnLc6 yielded mono- and difucosyl derivatives, whereas only a monofucosyl derivative was obtained when VI6NeuAcnLc6 was the acceptor. This is most probably due to different conformations at the terminus of the two acceptor gangliosides. The fucosyltransferase was incapable of transferring fucose to sialyl 2----3 lactotetraosylceramide (IV3NeuAcLc4).     

The widespread effect of beta 1,4-galactosyltransferase on N-glycan processing

We investigated beta 1,4-GalT (UDP-galactose: beta-d-N-acetylglucosaminide beta 1,4-galactosyltransferase) in terms of intracellular competition with GnT-IV (UDP-N-acetylglucosamine: alpha1,3-d-mannoside beta1,4-N-acetylglucosaminyltransferase) and GnT-V (UDP-N-acetylglucosamine: alpha1,6-d-mannoside beta 1,6-N-acetylglucosaminyltransferase). The beta 1,4-GalT-I gene was introduced into Chinese hamster ovary (CHO) cells producing human interferon (hIFN)-gamma (IM4/V/IV cells) and five clones expressing various levels of beta 1,4-GalT were isolated. As we previously reported, parental IM4/V/IV cells express high levels of GnT-IVa and -V and produce hIFN-gamma having primarily tetraantennary sugar chains. The branching of sugar chains on hIFN-gamma was suppressed in the beta 1,4-GalT-enhanced clones to a level corresponding to the intracellular activity of beta 1,4-GalT relative to GnTs. Moreover, the contents of hybrid-type and high-mannose-type sugar chains increased in these clones. The results showed that beta 1,4-GalT widely affects N-glycan processing by competing with GnT-IV, GnT-V, and alpha-mannosidase II in cells and also by some other mechanisms that suppress the conversion of high-mannose-type sugar chains to the hybrid type.     

The hypolipidemic compound cetaben induces changes in Golgi morphology and vesicle movement

The human hepatoma cell line HepG2 was used to study the effect of cetaben, a non-fibrate hypolipidemic drug, on cell morphology and vesicle distribution. Cetaben treatment correlated with a fragmentation and/or condensation of Golgi cisternae and the appearance of large electron-lucent vesicles. The Golgi apparatus, demonstrated, for example, by fluorescence-lectin histochemistry, was fragmented after cetaben treatment. The lectin-positive remnants were dispersed throughout the cytoplasm, but with a preference for being transported to tips of cells. However, microtubules and the intermediate filaments as well as the actin microfilaments were unchanged after cetaben treatment indicating that changes in Golgi morphology are not caused by alterations in the cytoskeleton. Cetaben decreases the cholesterol content due to inhibition of cholesterol biosynthesis. Changes in the intracellular cholesterol content are known to influence the intracellular vesicle distribution and are most likely responsible for cetaben-induced Golgi alterations, as depletion of cellular cholesterol by starvation or lovastatin and/or cyclodextrin treatment resulted in a similar redistribution of Golgi-derived wheat germ agglutinin vesicles. These lectin-stained vesicles colocalized with lysosomal marker proteins such as Limp-1 and Lamp-2, but not with the early endosomal markers Rab5 and EEA1. Upon removal of cetaben the lectin- and Limp-1/Lamp-2-costained vesicles dissociated and were transported back to the perinuclear region. Thus, cetaben-induced changes such as fragmentation of the Golgi apparatus and the dispersion of lysosomes away from their juxtanuclear location were reversible.Electronic Supplementary Material Supplementary material is available in the online version of this article at http://dx.doi.org/10.1007/s00418-004-0689-6     

Characterization of YIPF3 and YIPF4, cis-Golgi Localizing Yip domain family proteins

The Yip1 domain family (YIPF) proteins are homologues of yeast Yip1p and Yif1p, which are proposed to function in ER to Golgi transport. Here, we report the characterization of YIPF3 and YIPF4, homologues of human Yif1p and Yip1p, respectively. Immunofluorescence and immuno-electron microscopy showed that both YIPF3 and YIPF4 are clearly concentrated in the cis-Golgi. While YIPF4 was detected as a single mobility form consistent with its predicted molecular weight, three different mobility forms of YIPF3 were detected by western blotting. Biochemical and immunofluorescence experiments strongly indicated that YIPF3 is synthesized in the ER as a N-glycosylated form (40 kDa), is then O-glycosylated in the Golgi apparatus to become a lower mobility form (46 kDa) and finally becomes a higher mobility form cleaved at its C-terminal luminal domain (36 kDa). YIPF3 and YIPF4 form a complex in the Golgi apparatus, and this was suggested to be important for their proper localization and function. The knockdown of YIPF3 or YIPF4 in HeLa cells induced fragmentation of the Golgi apparatus, suggesting their involvement in the maintenance of the Golgi structure.     

An intrinsic membrane glycoprotein of the golgi apparatus with O-linked N-acetylglucosamine facing the cytosol

We have recently described the occurrence of integral membrane glycoproteins in rat liver smooth and rough endoplasmic reticulum with O-N-acetylglucosamine facing the cytosolic and luminal sides of the membrane (Abeijon, C., and Hirschberg, C. B. (1988) Proc. Natl. Acad. Sci. U.S.A. 85, 1010-1014). We now report that integral membrane glycoproteins with cytosolic facing O-N-acetylglucosamine also occur in membranes of rat liver Golgi apparatus. This was determined following incubation of vesicles from the Golgi apparatus, which were sealed and of the same membrane topographical orientation as in vivo, with UDP-[14C]galactose and saturating amounts of bovine milk galactosyltransferase. This enzyme does not enter the lumen of the vesicles and specifically catalyzes the addition of galactose, in a beta 1-4 linkage, to terminal N-acetylglucosamine. Under these conditions, galactose was transferred to a glycoprotein of molecular mass of 92 kDa. This protein was insoluble in sodium carbonate, pH 11.5, conditions under which integral membrane proteins remain membrane bound and was insensitive to treatment with peptide:N-glycosidase F. beta Elimination and chromatography showed that radiolabeled galactose was part of a disaccharide which was characterized as Gal beta 1-4GlcNAcitol. This glycoprotein is specific of the Golgi apparatus membrane. Intrinsic membrane glycoproteins with this unusual carbohydrate membrane orientation thus occur in the endoplasmic reticulum and Golgi apparatus of rat liver.     

Xylosylation and glucuronosylation reactions in rat liver Golgi apparatus and endoplasmic reticulum

We have studied in rat liver the subcellular sites and topography of xylosylation and galactosylation reactions occurring in the biosynthesis of the D-glucuronic acid-galactose-galactose-D-xylose linkage region of proteoglycans and of glucuronosylation reactions involved in both glycosaminoglycan biosynthesis and bile acid and bilirubin conjugation. The specific translocation rate of UDP-xylose into sealed, "right-side-out" vesicles from the Golgi apparatus was 2-5-fold higher than into sealed right-side-out vesicles from the rough endoplasmic reticulum (RER). Using the above vesicle preparations, we only detected endogenous acceptors for xylosylation in the Golgi apparatus-rich fraction. The specific activity of xylosyltransferase (using silk fibroin as exogenous acceptor) was 50-100-fold higher in Golgi apparatus membranes than in those from the RER. Previous studies had shown that UDP-galactose is translocated solely into vesicles from the Golgi apparatus. In these studies, we found the specific activity of galactosyltransferase I to be 40-140-fold higher in membranes from the Golgi apparatus than in those from the RER. The specific translocation rate of UDP-D-glucuronic acid into vesicles from the Golgi apparatus was 10-fold higher than into those from the RER, whereas the specific activity of glucuronosyltransferase (using chondroitin nonasaccharide as exogenous acceptor) was 12-30-fold higher in Golgi apparatus membranes than in those from the RER. Together, the above results strongly suggest that, in rat liver, the biosynthesis of the above-described proteoglycan linkage region occurs in the Golgi apparatus. The specific activity of glucuronosyltransferase, using bile acids and bilirubin as exogenous acceptor, was 10-25-fold higher in RER membranes than those from the Golgi apparatus. This suggests that transport of UDP-D-glucuronic acid into the RER lumen is not required for such reactions.     

Pulse-chase studies of the synthesis and intracellular transport of apolipoprotein B-100 in Hep G2 cells

The synthesis and secretion of apolipoprotein B-100 (apoB-100) have been studied in a human hepatoma cell line, the Hep G2 cells. The time needed for the synthesis of apoB-100 was estimated to be 14 min, which corresponds to a translation rate of approximately 6 amino acids/s. ApoB-100 was compared with albumin and alpha 2-macroglobulin as to the distribution between the membrane and the luminal content in the endoplasmic reticulum (ER) and the Golgi apparatus. The results suggested that apoB-100 approximately followed the distribution of these secretory proteins in the Golgi, while the ratios between the percent membrane-bound apoB-100 and percent membrane-bound albumin or alpha 2-macroglobulin were 3-4:1 in the ER. This may suggest that apoB-100 occurs in a membrane-associated form in ER prior to the integration in the lipoproteins. Pulse-chase studies combined with subcellular fractionation was used to investigate the kinetics for the intracellular transfer of apoB-100. A 3-min pulse of [35S]methionine was followed by an increase in apoB-100 radioactivity in the ER during the first 10-15 min of chase. The following 10-15 min of chase were characterized by linear decrease in apoB-100 radioactivity with a decay rate of approximately 6%/min. The residence kinetics for apoB-100 in the ER differed from that of transferrin and probably also from that of albumin. By comparing the time for the pulse maximum in ER with that in the denser Golgi fractions the time needed for the transfer between ER and Golgi could be estimated to be 10 min. The time needed for the secretion of newly synthesized apoB-100 was estimated to be 30 min. This indicates that the transfer of the protein through the Golgi apparatus to the extracellular space requires 20 min.