Patterns of cell differentiation revealed by L-[3H]fucose incorporation in Dictyostelium

A study of the incorporation of l-[6-³H]fucose and d-[6-³H]glucosamine hydrochloride was conducted during the development of the cellular slime mold Dictyostelium discoideum 1-H. Autoradiographs revealed that pulse-labeled vegetative amoebae incorporated [³H]fucose intracytoplasmically within 15 min. The majority of the cells had randomly scattered silver grains but the remainder were distinguished by a dense localized labeling which suggested that oligo or polysaccharide synthesis was occurring. The localized pattern of labeling attributed to active synthesis declines at aggregation and early conus formation. As the pseudoplasmodium makes the developmental transition from the conus to the culmination stages the localized pattern of [³H]fucose labeling was restricted to the prespore cells while the prestalk cells were devoid of label. Prespore vacuoles were not present at the onset of this transition and consequently [³H]fucose incorporation occurred in the cells prior to their differentiation into prespore cells. In contrast to cells composing earlier stages, mature spores exhibited [³H]fucose-containing substances at the cell surface. At appropriate stages certain cells actively synthesize slime and stalk sheath which were labeled with either [³H]fucose or [³H]glucosamine.Prestalk isolates were obtained by transecting migrating slugs. [³H]Fucose was incorporated within 10 min among the basal cells of the isolate in the localized pattern typically found in prespore cells. The incorporation of [³H]fucose occurred prior to prespore differentiation as certain preparations were devoid of prespore vacuoles. Prespore isolates differentiate prestalk cells which have lost the capacity to incorporate [³H]fucose. This investigation suggests that cell contacts and interactions may affect the incorporation of [³H]fucose.     

Rapidly metabolized glycoproteins in a neuroplastoma cell line

The metabolism of neuroblastoma cell glycoproteins was examined using l-[³H]fucose. Incubation of monolayer cultures with [³H]fucose resulted in a rapid uptake of the radioactive precursor and its incorporation into acid-insoluble macromolecules. Less than 3% of the [³H]fucose that was isolated from neuroblastoma cells by trichloroacetic acid precipitation was associated with glycolipids. The metabolism of fucosylated macromolecules was studied in cells which were labelled to a steady state, and then reincubated under conditions which limited reutilization of the radioactive precursor (40 mM unlabelled fucose). During reincubation of the cells, we observed a rapid metabolism (27% by 2 h)_ of the prelabelled macromolecules which stabilized within a cell generation time to give an overall rate of turnover of 9%. This rapid loss of radioactivity from the cells was not due to exocytosis since less than 4% of the [³H]-fucose was lost into the media as macromolecules during a 5 h reincubation period. The presence of 40 mM fucose in the media did not affect cell growth until after 24 h of incubation or cellular synthesis until after 15 h of incubation. When the metabolism of neuroblastoma cell glycoproteins was measured in the presence of 1.8 · 10^?4 M cycloheximide, there appeared to be a less rapid decrease in cell-associated specific activity, and an increased reutilization of [³H]fucose. Although the major proportion of the radioactivity remained as [³H]fucose, extensive incubation of neuroblastoma cells with this radioactive precursor led to increased amounts of tritium associated with other cellular components. However, a rapid rate of glycoprotein metabolism could also be demonstrated with cells incubated with [⁴C]fucose. This eliminated the possibility that the above results were restricted to the tritiated precursor and merely a reflection of hydrogen-tritium exchange.     

Biosynthesis and transport of glycoproteins in the small intestinal epithelium of rats: I. Developmental change and effect of hypophysectomy

The biosynthesis and intracellular transport of glycoproteins in duodenal absorptive cells of intact rats at 6 and 24 days and hypophysectomized rats at 24 days of age were studied after 20 min intralumenal pulse-labeling of d-[³H]galactose, l-[³H]fucose, or d-[³H]mannose. Autoradiographic studies showed that the incorporation of sugars increased significantly in intact rats between 6 and 24 days. When rats were hypophysectomized at 6 days of age, the intestinal epithelium at 24 days incorporated d-[³H]galactose at a level significantly lower than that of intact rats at 24 days. Hypophysectomy also interfered with the developmental increase in d-[³H]mannose, but not in l-[³H]fucose, incorporation. Biochemical study indicated that the radioactivity in the lipid-free acid-precipitable glycoproteins in the intestine of 24-day-old intact rats at 20 min after d-[³H]galactose injection was 129% and 97% higher than that in 6-day-old rats and in 24-day-old hypophysectomized rats, respectively. The patterns of intracellular transport of newly synthesized galactosylated or fucosylated glycoproteins in all animal groups were similar; the labeled glycoproteins were initially present in the Golgi and were transported through the smooth endoplasmic reticulum to either the lateral membrane or the brush-border membrane within 60 min after the injection of labeled sugars. The proportion of labeled glycoproteins that migrated to the brush-border membrane, however, increased about twofold in the intact rats between 6 and 24 days of age at 60–240 min after d-[³H]galactose injection. Hypophysectomy interfered with developmental increase in the transport of glycoproteins from the apical cytoplasm to the brush-border membrane. It was concluded that the incorporation of monosaccharide precursors into glycoproteins and the porportion of newly synthesized galactosylated or fucosylated glycoproteins transported to the brush-border membrane increase during postnatal development. The developmental changes are regulated, at least partially, by the pituitary gland.     

Separate binding sites in rat brain synaptic membranes for l-cysteine sulfinate and for l-glutamate

Binding of l-[³H]cysteine sulfinic acid (CSA) and l-[³H]glutamate were compared in various subcellular fractions and in the presence of a variety of pharmacological and ionic manipulations in order to test the possibility that the two amino acids possessed separate binding sites.The specific l-[³H]cysteine sulfinate binding was found to be enriched maximally in medium and high density synaptic membranes, while the crude mitochondrial synaptosomal fraction displayed the highest l-[³H]glutamate binding. The ratio of l-[³H]cysteine sulfinate binding/l-[³H]glutamate binding was variable across brain regions. Several compounds differentially affected l-[³H]cysteine sulfinate and l-[³H]glutamate binding. l-cysteine sulfinate was the most potent displacer regardless of the binding considered. Finally, while cations produced qualitatively similar effects on the binding of the two amino acids, quantitative differences were evident.In sum, these data revealed the complexity of l-[³H]cysteine sulfinate and l-[³H]glutamate binding. They suggest the existence of several binding sites and that some of these are shared by both substances. However, the results also indicate that separate binding sites for the two amino acids exist in synaptic membrane, giving further support to the hypothesis that cysteine sulfinate serves a neurotransmitter role in the central nervous system.     

Secreted proteins from rat Sertoli cells.

Sertoli cell cultures were prepared from the testes of 20-day-old rats. The proteins which were secreted by the cells into the culture medium were labeled with [³H]leucine or l-[³H]fucose. The proteins were concentrated by ultrafiltration and analysed by polyacrylamide slab gel electrophoresis (PAGE) in the presence of sodium dodecyl sulfate (SDS). Autofluorography of the gels at ?70 °C showed that the rat Sertoli cells synthesized and secreted at least 7 major polypeptides. The polypeptides had molecular weights ranging from 16 000 to 140 000 D. Proteins which were secreted from cultures of testicular fibroblasts and myoid cells had electrophoretic properties on SDS-PAGE which were different from Sertoli cell secreted proteins. Addition of FSH and testosterone to the Sertoli cell cultures increased the total synthesis and secretion of [³H]leucine-labeled proteins. No qualitative changes in the proteins as a result of hormone application could be detected. However, the synthesis of a polypeptide of molecular weight 48 000 was increased relative to the other secreted peptides if the cells were maintained in FSH and testosterone. The Sertoli cell secreted proteins were shown to be glycoproteins which can bind to ConA-Sepharose and can be labeled with [³H]fucose. Tunicamycin, a specific inhibitor of N-glycosylation, inhibited the secretion of [³H]proteins by 50% but had little effect on the intracellular protein synthesis.     

The metabolism of l[3-3H]lactate by isolated hamster liver cells

The rate of tritium removal from l[3-³H]lactate by hamster liver cells is faster than the analytical rate of lactate utilization, or the rate of ¹⁴C disappearance from l[U-¹⁴C]lactate, with the result that the ³H/¹⁴C ratio in residual lactate from l-[U-¹⁴C,3-³H]lactate decreases. However, addition of low concentrations (0.1 to 1.0 mM) of l-cycloserine, a glutamate pyruvate transaminase inhibitor, nearly equalizes the rates of isotope utilization from l-[3-³H]lactate and l-[U-¹⁴C]lactate. The results suggest a very limited rate of recycling of phosphoenolpyruvate back to pyruvate during gluconeogenesis from lactate in fasted hamster liver cells.     

The effect of tunicamycin on development of the mammalian embryonic pancreas

The role of cell-surface glycoproteins in histogenesis of the embryonic rat pancreas was investigated by studying the effect of tunicamycin (TM) on in vitro development. TM has been shown to block glycosylation of asparagine residues in glycoproteins by inhibiting formation of dolichol oligosaccharide intermediates. Exposure of Day 15 pancreatic rudiments to 1.0 μg TM/ml for 15 or 24 hr inhibited [³H]mannose, [³H]glucosamine, and [³H]fucose incorporation by 95, 85, and 90%, respectively, while [³H]leucine incorporation was reduced by 35%. Similar results were obtained with Day 17 rudiments. These trends were confirmed using sodium dodecyl sulfate-polyacrylamide slab gel electrophoresis and fluorography. Inhibition of [³H]monosaccharide incorporation correlated with reduced binding of RCA I-ferritin conjugates to the cell surface and both effects of TM were reversed by reculturing rudiments in medium lacking the antibiotic. Morphologically, TM treatment resulted in a delay in pancreatic histogenesis and this delay correlated with an inhibition of the normal increase in specific activity of amylase, an acinar cell secretory protein. These effects were not mimicked by treatment with cycloheximide at a concentration which inhibited [³H]leucine incorporation to the same degree observed with TM. The percentage of delayed rudiments decreased as reculturing in the absence of TM was extended.     

Distribution of radiolabeled l-glutamate and d-aspartate from blood into peripheral tissues in naive rats: Significance for brain neuroprotection

Excess l-glutamate (glutamate) levels in brain interstitial and cerebrospinal fluids (ISF and CSF, respectively) are the hallmark of several neurodegenerative conditions such as stroke, traumatic brain injury or amyotrophic lateral sclerosis. Its removal could prevent the glutamate excitotoxicity that causes long-lasting neurological deficits. As in previous studies, we have established the role of blood glutamate levels in brain neuroprotection, we have now investigated the contribution of the peripheral organs to the homeostasis of glutamate in blood. We have administered naive rats with intravenous injections of either l-[1-¹⁴C] Glutamic acid (l-[1-¹⁴C] Glu), l-[G-³H] Glutamic acid (l-[G-³H] Glu) or d-[2,3-³H] Aspartic acid (d-[2,3-³H] Asp), a non-metabolized analog of glutamate, and have followed their distribution into peripheral organs. We have observed that the decay of the radioactivity associated with l-[1-¹⁴C] Glu and l-[G-³H] Glu was faster than that associated with glutamate non-metabolized analog, d-[2,3-³H] Asp. l-[1-¹⁴C] Glu was subjected in blood to a rapid decarboxylation with the loss of ¹⁴CO₂. The three major sequestrating organs, serving as depots for the eliminated glutamate and/or its metabolites were skeletal muscle, liver and gut, contributing together 92% or 87% of total l-[U-¹⁴C] Glu or d-[2,3-³H] Asp radioactivity capture. l-[U-¹⁴C] Glu and d-[2,3-³H] Asp showed a different organ sequestration pattern. We conclude that glutamate is rapidly eliminated from the blood into peripheral tissues, mainly in non-metabolized form. The liver plays a central role in glutamate metabolism and serves as an origin for glutamate metabolites that redistribute into skeletal muscle and gut. The findings of this study suggest now that pharmacological manipulations that reduce the liver glutamate release rate or cause a boosting of the skeletal muscle glutamate pumping rate are likely to cause brain neuroprotection.     

AUTORADIOGRAPHIC STUDIES OF SYNTHESIS AND INTRACELLULAR MIGRATION OF GLYCOPROTEINS IN THE RAT ANTERIOR PITUITARY GLAND

The incorporation of [³H]fucose in the somatotrophic and gonadotrophic cells of the rat adenohypophysis has been studied by electron microscope autoradiography to determine the site of synthesis of glycoproteins and to follow the migration of newly synthesized glycoproteins. The pituitaries were fixed 5 min, 20 min, 1 h, and 4 h after the in vivo injection of [³H]fucose and autoradiographs analyzed quantitatively. At 5 min after [³H]fucose administration, 80–90% of the silver grains were localized over the Golgi apparatus in both somatotrophs and gonadotrophs. By 20 min, the Golgi apparatus was still labeled and some radioactivity appeared over granules. At 1 h and 4 h, silver grains were found predominantly over secretory granules. The kinetic analysis showed that in both protein-secreting cells (somatotrophs) and glycoprotein-secreting cells (gonadotrophs), the glycoproteins have their synthesis completed in the Golgi apparatus and migrate subsequently to the secretory granules. It is concluded from these in vivo studies that glycoproteins which are not hormones are utilized for the formation of the matrix and/or of the membrane of the secretory granules. The incorporation of [³H]fucose in gonadectomy cells (hyperstimulated gonadotrophs) was also studied in vitro after pulse labeling of pituitary fragments in medium containing [³H]fucose. The incorporation of [³H]fucose was localized in both the rough endoplasmic reticulum (ER) and the Golgi apparatus. Later, the radioactivity over granules increased while that over the Golgi apparatus decreased. The concentration of silver grains over the dilated cisternae of the rough ER was not found to be modified at the longest time intervals studied.     

Synthesis of glycoproteins in a single identified neuron of Aplysia californica

Incorporation of L-[³H]fucose into glycoproteins was studied in R2, the giant neuron in the abdominal ganglion of Aplysia. [³H]fucose injected directly into the cell body of R2 was readily incorporated into glycoproteins which, as shown by autoradiography, were confined almost entirely to the injected neuron. Within 4 h after injection, 67% of the radioactivity in R2 had been incorporated into glycoproteins; at least 95% of these could be sedimented by centrifugation at 105,000 g, suggesting that they are associated with membranes. Extraction of the particulate fraction with sodium dodecyl sulfate (SDS), followed by gel filtration on Sephadex G-200 and polyacrylamide gel electrophoresis in SDS revealed the presence of only five major radioactive glycoprotein components which ranged in apparent molecular weight from 100,000 to 200,000 daltons. Similar results were obtained after intrasomatic injection of [³H]N-acetylgalactosamine. Mild acid hydrolysis of particulate fractions released all of the radioactivity in the form of fucose. When ganglia were incubated in the presence of [³H]fucose, radioactivity was preferentially incorporated into glial cells and connective tissue. In contrast to the relatively simple electrophoretic patterns obtained from cells injected with [³H]fucose, gel profiles of particulate fractions labeled with [¹⁴C]valine were much more complex.     

The removal of cell surface material by enzymes used to dissociate mammary-gland cells

Summary Treatment of mouse mammary epithelial cells (MMEC) with various enzymes used for dispersing and transferring cells results in extensive digestion of materials on the cell surfaces. MMEC biosynthetially labeled with [³H]fucose, [¹⁴C]fucose and [³H]amino acids or with¹²⁵I by the lactoperoxidase method were exposed to either collagenase plus hyaluronidase, followed by pronase, or to trypsin in concentrations and conditions currently used for cell dispersion. Whereas the latter enzyme preparation solubilized 76% of the trichloroacetic acid precipitable radioactive fucose and 96% of the protein-bound¹²⁵I, collagenase plus hyaluronidase treatment released lesser amounts of each label. Subsequent treatment of the cells with pronase removed additional surface-labeled materials, but the total amounts released were still less than when the trypsin preparation alone was employed. Released cell surface materials were analyzed by gel chromatography. Some of the peaks obtained also were examined by polyacrylamide gel electrophoresis. The labeled materials that remained attached to the MMEC after enzymatic treatment were investigated by these two methods as well. We could show that collagenase plus hyaluronidase solubilized three main glycoprotein components from the cell surface. In addition, we could show that the extensive cell surface damage caused by these two enzyme preparations was due to the high proteolytic activity present in these preparations as judged by their ability to hydrolyze rabbit gamma globulin labeled with¹²⁵I. Even though their membranes were extensively damaged by the enzyme treatments, the dispersed cells could be cultured successfully in vitro and could incorporated fucose into their surfaces in a manner similar to that by intact tissue. Through the use of gel-filtration (cochromatography of [¹⁴C]fucose and [³H]fucose cell surface materials), we could demonstrate the identity of cell surface glycoproteins synthesized by cultured cells and by intact tissue. This work was supported by Grant Nos. CA 11736 and CA 19455 from the National Cancer Institute, and Biomedical Research Support Grant No. RR05467 from the National Institutes of Health, DHEW.     

Incorporation of fucose into HeLa cell plasma membranes during the cell cycle

The metabolism of HeLa cell plasma membranes during the cell cycle was studied by following the incorporation of radioactive precursor l-[³H]fucose into plasma membranes of synchronized cells. Maximal incorporation of the radioactive precursor was observed in late S phase of the cell cycle. This discrete period of increased incorporation of precursor into the plasma membranes implies the existence of a distinct control mechanism which may relate cell surface phenomena to the cell cycle.     

A comparison of [3H]galactose and [3H]fucose uptake with the morphological and histochemical changes observed in mucous secretion in chemically induced rat colonic carcinoma

Summary The alterations in carbohydrate metabolism which occur in the distal colon of rats during carcinogenesis induced by dimethylhydrazine were investigated using [³H]galactose and [³H]fucose as glycoprotein precursors.A statistically significant decrease in [³H]galactose uptake was observed in dysplastic epithelia. These findings are consistent with the alterations in mucin composition with predominance of sialomucins shown in these areas by histochemical methods. Furthermore, changes in the gradient of [³H]galactose incorporation along the crypt epithelium were also found in the histological and histochemically non-involved colonic mucosa of dimethylhydrazine-treated rats, as compared with controls.No significant variations were seen in [³H]fucose incorporation.These results correlate well with our previous histochemical observations and are further evidence of the profound alterations in glycoprotein synthesis affecting the whole colonic mucosa during carcinogenesis.     

KINETICS OF ENTRY OF PROTEINS INTO THE MYELIN MEMBRANE1

Brain slices were prepared from 17-day old rats, and incubated with [³H]glycine or [³H]-leucine to label proteins. Myelin was isolated from the slices, and the proteins were separated by discontinuous gel electrophoresis in buffers containing sodium dodecyl sulfate. Radioactive basic and Wolfgram proteins appeared in myelin at similar initial rates, and their entry was nearly linear between 15 and 120 min with no detectable lag. Radioactive proteolipid protein appeared in myelin at one-fourth the rate of the basic and Wolfgram proteins between 0 and 30 min, then entered at a rate comparable to the other proteins between 45 and 120 min. When cycloheximide (0.2 mM) or puromycin (1.0 mM) was added, appearance of newly labeled basic and Wolfgram proteins in myelin stopped while proteolipid protein continued to appear in myelin at a normal rate for at least 30 min. Chase experiments with unlabeled glycine had similar effects. These results indicate the existence of a previously synthesized precursor pool of proteolipid protein with a 30-min interval between synthesis of proteolipid protein and its appearance in myelin. Incorporation of [³H]fucose into glycoprotein of the myelin sheath was studied, as was inhibition of incorporation of radioactivity by the use of either cycloheximide, or dilution with unlabeled fucose. The results indicated fucosylation of a sizable pool of presynthesized protein and a delay of 30 min between fucosylation of these polypeptides and their subsequent appearance in myelin as glycoproteins.     

Axonal transport of newly synthesized glycoproteins in a single identified neuron of Aplysia californica

Increasing amounts of glycoprotein synthesized from L-[³H]fucose injected into the cell body of R2, an identified Aplysia neuron, were found in the right pleuro-abdominal connective. Autoradiography revealed that the glycoproteins were localized in the axon of R2. Glycoproteins appearing in the axon presumably were synthesized in the cell body, since no significant incorporation was observed when [³H]fucose was injected directly into the axon. [³H]glycoproteins were detected in the connective after a delay of 1 h after intrasomatic injection. Thereafter, transport from the cell body was rapid, and by 10 h after injection, 45% of the total neuronal [³H]glycoprotein had appeared in the axon. By analysing the radioactivity in cell body and connective 4, 10, and 15 h after injection, we found that [³H]glycoproteins were transported selectively compared to nonmacromolecular material. Sequential sectioning of the connective revealed that [³H]glycoproteins were transported in discrete waves. The population of membrane-associated [³H]glycoproteins in the axon differed from that in the cell body. Two of the five somatic components appeared to be transported preferentially. In addition a new component appeared in the axon 10 h after injection.     

Respiration of 14CO2 by intact animals of various species given l-[1-14C]fucose or d-[1-14C]arabinose

The production of ¹⁴CO₂ from l-[1-¹⁴C]fucose and d-[1-¹⁴C]arabinose has been studied in five mammalian species.Cats, guinea pigs, mice, and rabbits respired about 22% of the label of l[1-¹⁴C]fucose or of d-[1-¹⁴C]arabinose within 6 h after intraperitoneal injection of the sugar. Rats respired only 1.5% of the l-fucose label and 5% of the d-arabinose label in the same time period.Liver homogenates from cat, guinea pig, and rabbit produced significantly more ¹⁴CO₂ from l-[1-¹⁴C]fucose or d-[1-¹⁴C]arabinose than mouse or rat liver homogenates. Unlike those of the other species, guinea pig liver homogenates had very low l-fucose dehydrogenase activity.The results suggest that substantial catabolism of l-fucose and d-arabinose occurs in the tissues of some animal species. Investigators wishing to employ l-fucose as a tracer of glycoprotein metabolism must, therefore, ensure that the species that they employ does not metabolize l-fucose to products interfering with their studies.     

Biosynthesis and release of glycoproteins by human skin fibroblasts in culture

1. Confluent human skin fibroblasts maintained in a chemically defined medium incorporate l-[1-³H]fucose in a linear manner with time into non-diffusible macromolecules for up to 48h. Chromatographic analysis demonstrated that virtually all the macromolecule-associated ³H was present as [³H]fucose. 2. Equilibrium CsCl-density-gradient centrifugation established that [³H]fucose-labelled macromolecules released into the medium were predominantly glycoproteins. Confirmation of this finding was provided by molecular-size analyses of the [³H]fucose-labelled material before and after trypsin digestion. 3. The [³H]fucose-labelled glycoproteins released into fibroblast culture medium were analysed by gel-filtration chromatography and sodium dodecyl sulphate/polyacrylamide-gel electrophoresis. These techniques demonstrated that the major fucosylated glycoprotein had an apparent mol.wt. of 230000–250000; several minor labelled species were also detected. 4. Dual-labelling experiments with [³H]fucose and ¹⁴C-labelled amino acids indicated that the major fucosylated glycoprotein was synthesized de novo by cultured fibroblasts. The non-collagenous nature of this glycoprotein was established by three independent methods. 5. Gel-filtration analysis before and after reduction with dithiothreitol showed that the major glycoprotein occurs as a disulphide-bonded dimer when analysed under denaturing conditions. Further experiments demonstrated that this glycoprotein was the predominant labelled species released into the medium when fibroblasts were incubated with [³⁵S]cysteine. 6. The relationship between the major fucosylated glycoprotein and a glycoprotein, or group of glycoproteins, variously known as fibronectin, LETS protein, cell-surface protein etc., is discussed.     

[3 5S]Sulfate incorporation into myelin glycoproteins II. Peripheral nervous tissue

The in vivo incorporation of [^{3 5}S]sulfate, [³H]fucose and [³H]leucine into sciatic nerve myelin was investigated. Polyacrylamide gel electrophoresis of the proteins indicated that the ^{3 5}S-labeling of proteins occurred almost exclusively in the major myelin protein. A smaller myelin glycoprotein migrating just ahead of the major one was labeled with [³H]fucose but did not incorporate ^{3 5}S to a detectable extent. There was little or no ^{3 5}S associated with basic proteins on polyacrylamide gels when the proteins were extracted with chloroform/methanol. Fucose-labeled myelin glycoproteins were converted to glycopeptides by pronase digestion. The glycopeptides gave a single peak on Sephadex G-50 in which the ³H and ^{3 5}S coincided. The association of ^{3 5}S with glycopeptides was not caused by binding of sulfatide or free inorganic sulfate. This study shows that the major myelin protein in the sciatic nerve of the rat is glycosylated and sulfated.     

Incorporation of fucose into the carbohydrate moiety of phytohemagglutinin in developing Phaseolus vulgaris cotyledons

Incubation of developing cotyledons of P. vulgaris with [³H]fucose resulted in the incorporation of radioactivity into the cell wall, membranous organelles and soluble macromolecules. Fractionation of the proteins by sodium dodecyl sulfate-polyacrylamide gel electrophoresis, followed by fluorography, showed that phytohemagglutinin (PHA) was the major fucosylated protein synthesized in the cotyledons. Incorporation of fucose into PHA occurred in the membranous organelle fraction, and the radioactive fucose remained associated with the PHA during a 20-h chase of the radioactivity. Tunicamycin inhibited the incorporation of glucosamine and fucose into PHA to the same extent (65%), indicating the involvement of a lipid intermediate in the incorporation of fucose, or the attachment of fucose to the high-mannose oligosaccharide moiety of newly synthesized PHA. Digestion with proteinase K of [³H]fucose- or [³H]glucosamine-labeled PHA resulted in the formation of glycopeptides of similar size. These glycopeptides were partially resistant to digestion with endo-β-N-acetylglucosaminidase H, even after the removal of fucose by mild acid hydrolysis. We postulate, on the basis of these experiments, that the transport of PHA from the endoplasmic reticulum to the protein bodies is accompanied by the modification of its oligosaccharide side-chain. This modification involves inter alia the attachment of fucose, and renders the oligosaccharide side-chain resistant to digestion with endo-β-N-acetylglucosaminidase H. Analogy with animal glycoproteins indicates that this modification probably occurs in the Golgi apparatus.