Ectogalactosyltransferase. Presence of enzyme and acceptors on the rat lymphocyte cell surface.

Experiments are described to demonstrate the existence of ectogalactosyltransferase activity on the lymphocyte surface. The procedures described enable us to exclude the possibility of misleading results due to precursor hydrolysis and intracellular utilization of the free galactose. This depicted transferase is able to catalyse the transfer of a galactosyl residue from UDP-galactose to a nonphagocytosable exogenous acceptor and to endogenous membrane acceptors. The cells galactosylated in this way acquired new agglutinating properties with soybean agglutinin, which proves the external position of the galactosyl residues incorporated on the cell surface.     

Mouse lymphoma cell lines resistant to pea lectin are defective in fucose metabolism

Two mutants of the BW5147 mouse lymphoma cell line have been selected for their resistance to the toxic effects of pea lectin. These cell lines, termed PLR1.3 and PHAR1.8 PLR7.2, have a decreased number of high affinity pea lectin-binding sites (Trowbridge, I.S., Hyman, R., Ferson, T., and Mazauskas, C. (1978) Eur. J. Immunol. 8, 716-723). Intact cell labeling experiments using [2-3H]mannose indicated that PLR1.3 cells have a block in the conversion of GDP-[3H]mannose to GDP-[3H]fucose whereas PHAR1.8 PLR7.2 cells appear to be blocked in the transfer of fucose from GDP-[3H]fucose to glycoprotein acceptors. In vitro experiments with extracts of PLR1.3 cells confirmed the failure to convert GDP-mannose to GDP-fucose and indicated that the defect is in GDP-mannose 4,6-dehydratase (EC 4.2.1.47), the first enzyme in the conversion of GDP-mannose to GDP-fucose. The block in the PLR1.3 cells could be bypassed by growing the cells in the presence of fucose, demonstrating that an alternate pathway for the production of GDP-fucose presumably via fucose 1-phosphate is functional in this line. PLR1.3 cells grown in 10 mM fucose showed normal high affinity pea lectin binding. PHRA1.8 PLR7.2 cells synthesize GDP-fucose and have normal or increased levels of GDP-fucose:glycoprotein fucosyltransferase when assayed in vitro. The fucosyltransferases of this clone can utilize its own glycoproteins as fucose acceptors in in vitro assays. These findings indicate that this cell line fails to carry out the fucosyltransferase reaction in vivo despite the fact that it possesses the appropriate nucleotide sugar, glycoprotein acceptors, and fucosyltransferase. The finding of decreased glycoprotein fucose in two independent isolates of pea lectin-resistant cell lines and the restoration of high affinity pea lectin binding to PLR1.3 cells following fucose feeding strongly implicates fucose as a major determinant of pea lectin binding.     

Acceptor specificity and tissue distribution of three human alpha-3-fucosyltransferases

Based on the capacity to transfer alpha-L-fucose onto type-1 and type-2 synthetic blood group H and sialylated acceptors, a comparison of the alpha-3-fucosyltransferase activities of different human tissues is shown. Three distinct acceptor specificity patterns are described: (I) myeloid alpha-3-fucosyltransferase pattern, in which leukocytes and brain enzymes transfer fucose actively onto H type-2 acceptor and poorly onto sialylated N-acetyllactosamine: (II) plasma alpha-3-fucosyltransferase (EC 2.4.1.152), in which plasma and hepatocyte enzymes transfer, in addition, onto the sialylated N-acetyllactosamine; (III) Lewis alpha-3 4-fucosyltransferase (EC 2.4.1.65), in which gall-bladder kidney and milk enzymes transfer, in addition, onto type-1 acceptors. The small amount (less than 10%) of alpha-3-fucosyltransferase activity found in the plasma of an alpha-3-fucosyltransferase-deficient individual had a myeloid-type acceptor pattern, suggesting that this small proportion of the plasma enzyme is derived from leukocytes. In addition to the three acceptor specificity patterns, these enzyme activities can be differentiated by their optimum pH: 8.0-8.7 for the enzymes from myeloid cells and brain. 7.2-8.0 for liver enzymes and 6.0-7.2 for gallbladder enzymes. Milk samples had two alpha-3-fucosyltransferase activities, the Lewis or alpha-3/4-fucosyltransferase under control of the Lewis gene and an alpha-3-fucosyltransferase with plasma acceptor pattern which was independent of the control of the Lewis gene. The apparent affinity for GDP-fucose of the myeloid-like enzyme was weaker than those of the plasma and Lewis-like enzymes. The apparent affinities for H type 2 and sialylated N-acetyllactosamine were stronger for exocrine secretions as compared to the plasma and myeloid enzymes. The plasma type of alpha-3-fucosyltransferase activity was more sensitive to N-ethylmaleimide and heat inactivation than the samples with myeloid-like alpha-3-fucosyltransferase activity.     

Biosynthesis of polyuronides inMucor rouxii: Partial characterization of fucosyl transferase

Most of the fucosyl transferase activity fromMucor rouxii was detected in a crude membrane fraction. The enzyme transferredl-fucose from GDP-fucose to endogenous and exogenous acceptors. When crude membrane fractions were treated with neutral detergents such as Trition X-100 or Brij 36 T enzyme activity became dependent on exogenous acceptors such as mucoric acid or mucoran. Brij-treated membrane fractions showed maximum fucosyl transferase activity at pH 6.5, and at a temperature between 22 and 28°C. The cations Mn²⁺, Mg²⁺, Co²⁺, Zn²⁺, Fe²⁺, and Ca²⁺ activated the enzyme about twofold. The former was slightly more stimulatory at 4 mM. Km for GDP-fucose was 10 μM. Evidence was obtained that mucoric acid serves as acceptor for fucosyl moieties. Acid hydrolysis of the product synthesized from GDP-fuc by Brij-treated membrane fractions revealed fucose as the major radioactive sugar.     

Ectoglycosyltransferase activity in suspensions and monolayers of cultured fibroblasts.

Fibroblasts suspended by a brief exposure to EDTA have the ability to transfer the carbohydrate moiety of exogenous nucleotide-sugars to endogenous acceptors. Monolayers of the same cells do not have this ability. Both suspensions and monolayers can transfer carbohydrate to exogenous glycose acceptors. The cells can glycosylate exogenous desialized, β-glactosidase treated fetuin utilizing either UDP-[¹⁴C]-galactose as a direct donor or [³H]-galactose as a precursor to a glycose donor.     

Cell surface glycosyl transferase activities in liver cells of developing chicken embryos

Summary The cell surface of embryonic chick liver cells contains transferases for mannose, fucose, galactose, N-acetyl-glucosamine and N-acetyl-neuraminic acid. Liver cells obtained by trypsin-dissociation of the tissue use the corresponding exogenous sugar nucleotides as substrates. The activities of the enzymes tested do not depend neither on the dissociation procedure nor onde novo protcin synthesis. They vary considerably during development of the embryos, reaching maximal values at the 8th ± 1 day and at the 12th ± 1 day. Glycoproteins are the final stable endogenous acceptors for all sugars. Mannose transfer proceeds via a two or multistep reaction sequence. In a first step labile lipophilic intermediates are formed. Mannose can be liberated by treating the intermediates with 0.1n HCI at 100°C. In a second reaction step mannose becomes attached to glycoproteins. From embryonic chick liver cells a glycopeptide fraction has been obtained by pronase digestion followed by several purification steps. The purified glycopeptides inhibit all transferase systems and act as exogenous acceptors for mannose transfered from exogenous GDP-mannose.     

Characterization of a cytosolic fucosylation pathway in Dictyostelium.

FP21 is a 21-kDa fucoprotein which fractionates with the cytosol after high-speed centrifugation of gently lysed Dictyostelium cells. Less than 0.7% of FP21 is associated with vesicles. In proliferating cells, 4 x 10(5) fucosyl moieties/cell are associated with FP21 as anionic, possibly O-linked oligosaccharides equal in size to 4.8 glucose units. FP21 is underfucosylated in a mutant strain (HL250) that depends on extracellular fucose for synthesis of GDP-fucose. To determine the cellular site of FP21 fucosylation, cytosolic and vesicular preparations from strain HL250 were compared for their ability to transfer fucose from GDP-fucose to FP21. Cytosolic preparations fucosylate endogenous FP21 in a time-, concentration-, and divalent cation-dependent fashion, with a Km for GDP-fucose of 1.4 microM. Activity in normal cell cytosol is dependent on exogenous mutant FP21, demonstrating that FP21 is normally fully fucosylated. Both mutant and normal cytosols are also able to alpha-fucosylate a type 1 glycolipid substrate (8-methoxycarbonyloctyl-Gal beta 1-3 beta GlcNAc), but not related substrates, with Km values for the type 1 glycolipid of 0.99 mM and for GDP-fucose of 1.6 microM. Competitive inhibition between FP21 and the type 1 glycolipid shows that the same enzyme fucosylates both substrates. Intact and permeabilized vesicle preparations from wild-type cells are unable to fucosylate FP21 or the type 1 glycolipid by a divalent cation-dependent mechanism, and thus are devoid of FP21-fucosyltransferase. Since control experiments showed that vesicle leakage is minimal during cytosol preparation, these results indicate that FP21 is synthesized and fucosylated in the cytosolic compartment, by an unusual soluble fucosyltransferase.     

Localization and modulation of galactosyltransferase during in vitro proliferation of a human ovarian adenocarcinoma cell line

A cell line, IGROV1, originating from a human ovarian cancer, releases a galactosyltransferase activity in its culture medium during proliferation. The proliferating IGROV1 cells appear as two populations: some cells grow in floating clusters whereas the greater part of them adhere to the culture substrate. The study of galactose transfer by intact cells onto an exogenous glycoprotein acceptor (ovomucoid) shows the presence of surface-associated galactosyltransferase onto the two cellular sub-populations. Opposite to intracellular activity, surface-associated and released galactosyltransferase activities depend on cellular adhesion and proliferation.     

Incorporation of fucose into glycans independent of the GDP-fucose transporter SLC35C1 preferentially utilizes salvaged over de novo GDP-fucose

Mutations in the SLC35C1 gene encoding the Golgi GDP-fucose transporter are known to cause leukocyte adhesion deficiency II. However, improvement of fucosylation in leukocyte adhesion deficiency II patients treated with exogenous fucose suggests the existence of an SLC35C1-independent route of GDP-fucose transport, which remains a mystery. To investigate this phenomenon, we developed and characterized a human cell–based model deficient in SLC35C1 activity. The resulting cells were cultured in the presence/absence of exogenous fucose and mannose, followed by examination of fucosylation potential and nucleotide sugar levels. We found that cells displayed low but detectable levels of fucosylation in the absence of SLC35C1. Strikingly, we show that defects in fucosylation were almost completely reversed upon treatment with millimolar concentrations of fucose. Furthermore, we show that even if fucose was supplemented at nanomolar concentrations, it was still incorporated into glycans by these knockout cells. We also found that the SLC35C1-independent transport preferentially utilized GDP-fucose from the salvage pathway over the de novo biogenesis pathway as a source of this substrate. Taken together, our results imply that the Golgi systems of GDP-fucose transport discriminate between substrate pools obtained from different metabolic pathways, which suggests a functional connection between nucleotide sugar transporters and nucleotide sugar synthases.     

Glycosylation of proteins from sugar nucleotides by whole cells. Effect of ammonium chloride treatment on mouse thymocytes.

When thymocytes are treated with iso-osmotic NH4Cl, the sugar incorporation into endogenous acceptors from labelled sugar nucleotides is largely increased compared with that in control thymocytes. This effect was obtained with labelled GDP-mannose, UDP-galactose and CMP-N-acetylneuraminic acid. The stimulation observed with NH4Cl-treated thymocytes does not involve the glycosylation of exogenous acceptors, and it was proved that the NH4Cl treatment (1) does not stimulate glycosyltransferase activities themselves, (2) does not lead to the release of soluble glycosyltransferases as the result of an extensive lysis of the thymocytes and (3) does not cause the emergence of glycosyltransferases at the cell surface. In fact, electron-microscopy observations showed that, although marked changes had occurred in the cytoplasm, the plasma membrane is sufficiently maintained to allow the cell to keep roughly its original shape and to retain the intracellular vesicles. We thus demonstrate that this stimulation is due to an enhancement of the entry of sugar nucleotides into the cell. As demonstrated by the inclusion of Trypan Blue within the cells, and the non-stimulation of glycosylation of exogenous large-molecular-mass acceptors, the effect of NH4Cl seems to be limited to the penetration of small-molecular-sized compounds through the plasma membrane. Thus NH4Cl treatment allows the labelled sugar nucleotides to penetrate the cell and to behave as the cellular pool to be utilized for glycosylation by intracellular vesicles.     

UDP-N-acetylglucosamine-glycoprotein N-acetylglucosaminyltransferase in regenerating rat liver

The assay condition for N-acetylglucosaminyltransferase activities in rat liver microsomal fraction was developed. The enzyme activities towards endogenous acceptors within 48 h after partial hepatectomy were lower than in controls, exceeding the control level by 96 h, and then higher than in controls up to 240 h after the operation. The changes in N-acetylglucosaminyltransferase activities towards exogenous acceptor (UDP-2-acetamido-2-deoxy-D-glucose: glycoprotein 2-acetamido-2-deoxy-D-glucosyltransferase, EC 2.4.1.51) were consistent with those in the enzyme activities towards endogenous acceptors at 144 h, but not at 48 h, after the operation. The contents of protein and the levels of protein-bound hexosamine in the liver microsomes were decreased at early period of liver regeneration. These results suggest that the acceptor capacity of liver microsomal proteins is diminished during first 48 h of the regeneration. This may be responsible for the decreased transfer of the amino sugar to nascent glycoproteins. However, the enzyme activity was enhanced at 144 h and the level of endogenous acceptors may increase.     

Amino acid transport and metabolism in nitrogen-starved cells of Saccharomyces cerevisiae.

Nitrogen-starved yeast derepress a general amino acid permease which transports basic and hydrophobic amino acids. Although both groups of amino acids are metabolized, the derivatives of the basic amino acids are retained by the cells, whereas those of the hydrophobic amino acids are released as acidic and neutral deaminated derivatives. The release of the deaminated derivatives of the hydrophobic amino acids only occurs in the presence of glucose, which presumably produces amino acceptors. The accumulation of intracellular amino acids results in trans-inhibition of the uptake of exogenous amino acids whether the intracellular amino acid is a basic amino acid or the product of intracellular transamination from a hydrophobic amino acid. Variation of permease and transaminase activity was measured during growth under repressed (ammonia-grown) and derepressed (proline-grown) conditions. Maximum levels for both activities occurs at the mid-exponential phase.     

Co2+ is able to substitute for Mn2+ in some exogenous and endogenous galactosyltransferase reactions

The ability of Co2+ to substitute for Mn2+ in exogenous and endogenous galactosyltransferase reactions was tested. Exogenous transfer was measured towards different high and low molecular weight galactose acceptors using galactosyltransferase from the following sources: crude serum, the serum enzyme partially purified by affinity chromatography and a pure enzyme preparation from milk. Endogenous transfer was estimated in preparations from human urinary bladder tumor cells and from rat liver microsomal fractions. The results show that Co2+ is able to substitute for Mn2+ in some exogenous and endogenous galactosyltransferase reactions. This ability seems to depend on the molecular structure of the galactose acceptor as well as on the nature of the enzyme.     

The GDP-fucose:N-acetylglucosaminide 3-alpha-L-fucosyltransferases of LEC11 and LEC12 Chinese hamster ovary mutants exhibit novel specificities for glycolipid substrates

Previous studies have shown that the GDP-fucose:N-acetylglucosaminide 3-alpha-L-fucosyltransferase (alpha (1,3) fucosyltransferase (Fuc-T)) activities expressed by the Chinese hamster ovary cell mutants LEC11 (Fuc-TI) and LEC12 (Fuc-TII) are different enzymes and indicated that Fuc-TI might act on sialylated lactosamine sequences (Campbell, C., and Stanley, P. (1984) J. Biol. Chem. 259, 11208-11214). In this paper we show that CSLEX-1, a monoclonal antibody specific for NeuNac alpha (2,3)Gal beta (1,4)(Fuc alpha (1,3))GlcNAc beta 1 sequences, bound to LEC11 cells but not to LEC12 cells. Direct evidence that Fuc-TI could act on sialylated substrates was sought with a series of glycolipid acceptors. Optimal assay conditions in crude cell extracts were determined with nLc4, a glycolipid which accepted fucose with both Fuc-TI and Fuc-TII to generate the Lex antigenic determinant. The two enzymes differed in their detergent sensitivities, pH optima, Mn2+ requirements, and apparent Km values for nLc4. When sialylated glycolipids were examined as substrates, Fuc-TI added fucose to IV3NeuNAcnLc4 but not to IV6NeuNAcnLc4, whereas Fuc-TII was unable to utilize either glycolipid as a substrate. Further studies showed that Fuc-TI and Fuc-TII possess novel specificities for glycolipids containing two lactosamine sequences as potential fucose acceptors. Fuc-TI exhibited good activities with VI3NeuNAcnLc6 and VI6NeuNAcnLc6 whereas Fuc-TII had very low activity with both substrates. Glycosidase digestions of the labeled products showed that Fuc-TI added fucose primarily to the internal N-acetylglucosamine of both glycolipids. The same preference for the internal N-acetylglucosamine was shown by Fuc-TI when nLc6 was the acceptor. In contrast, Fuc-TII preferred to transfer fucose to the external acceptor site of nLc6, consistent with the low activities of Fuc-TII with sialylated nLc6 derivatives. Thus the two enzymes preferentially add fucose to different N-acetylglucosamines in the same substrate, nLc6. This indicates that the biosynthetic pathway for fucosylation of polylactosamine sequences in glycolipids and glycoproteins will vary depending upon the particular alpha (1,3)fucosyltransferase present.     

UDP-N-acetylglucosamine-glycoproteinN-acetylglucosaminyltransferase in regenerating rat liver

The assay condition for N-acetylglucosaminyltransferase activities in rat liver microsomal fraction was developed. The enzyme activities towards endogenous acceptors within 48 h after partial hepatectomy were lower than in controls, exceeding the control level by 96 h, and then higher than in controls up to 240 h after the operation. The changes in N-acetylglucosaminyltransferase activities towards exogenous acceptor (UPD-2-acetamido-2-deoxy-D-glucose: glycoprotein 2-acetamido-2-deoxy-D-glucosyltransferase, EC 2.4.1.51) were consistent with those in the enzyme activities towards endogenous acceptors at 144 h, but not at 48 h, after the operation. The contents of protein and the levels of protein-bound hexosamine in the liver microsomes were decreased at early period of liver regeneration.These results suggest that the acceptor capacity of liver microsomal proteins is diminished during first 48 h of the regeneration. This may be responsible for the decreased transfer of the amino sugar to nascent glycoproteins. However, the enzyme activity was enhanced at 144 h and the level of endogenous acceptors may increase.     

Fucosylation of exogenous xyloglucans by pea microsomal membranes

Microsomal membrane preparations from growing regions of etiolated pea stems catalyzed the transfer of [14C]fucosyl units from GDP-[U-14C]-L-fucose into exogenously added xyloglucan acceptors, as well as into endogenous xyloglucan. The transfer was more effective using nonfucosylated tamarind seed xyloglucan than with pea wall xyloglucan in which almost all galactose units are already fucosylated. Hydrolysis of products by endo-1,4-beta-D-glucanase yielded in each case radioactive nonasaccharide as the main fucosylated product. UDP-galactose enhanced the fucosylation of endogenous primer but it had little effect on fucosyl transfer to exogenously added xyloglucans. Low-molecular-weight nonfucosylated oligosaccharide fragments up to the octasaccharide Glc4Xyl3Gal (obtained by endoglucanase action on tamarind seed xyloglucan) were ineffectual as fucosyl acceptors but inhibited the fucosylation of endogenous as well as of added xyloglucan. With octasaccharide, the inhibition was competitive in relation to the xyloglucan acceptor (Ki = 70 microM) and noncompetitive in relation to the donor GDP-fucose (Ki = 210 microM). It is concluded that fucosyltransferase acts independently and in a noncoordinated manner from other glycosyltransferases that are required to synthesize xyloglucan. Its active site recognizes a fragment longer than the galactosylated octasaccharide unit before transfucosylation will ensue.     

EVIDENCE FOR CELL-SURFACE GLYCOSYLTRANSFERASES : Their Potential Role in Cellular Recognition

Intact chicken embryo neural retina cells have been shown to catalyze the transfer of galactose-¹⁴C from uridine diphosphate galactose (UDP-galactose) to endogenous acceptors of high molecular weight as well as to exogenous acceptors. Four lines of evidence indicate that the galactosyltransferases catalyzing these reactions are at least partly located on the outside surface of the plasma membrane: (a) there is no evidence for appreciable uptake of sugar-nucleotides by vertebrate cells nor did unlabeled galactose, galactose 1-phosphate, or UDP-glucose interfere with the radioactivity incorporated during the reaction; (b) the cells remained essentially intact during the course of the reaction; (c) there was insufficient galactosyltransferase activity in the cell supernatants to account for the incorporation of galactose-¹⁴C into cell pellets; and (d) the intact cells could transfer galactose to acceptors of 10⁶ daltons, and the product of this reaction was in the extracellular fluid. Appropriate galactosyl acceptors interfered with the adhesive specificity of neural retina cells; other compounds, which were not acceptors, had no effect. These results suggested that the transferase-acceptor complex may play a role in cellular recognition.     

A collagen:glucosyltransferase at the surface of malignant fibroblasts

3T12 fibroblasts possess glucosyltransferases that catalyze the transfer of glucose from UDP-Glucose to galactosylhydroxylysyl residues on collagenous acceptors. The presence of the enzyme activity at the cell surface is indicated by the following findings: a) suspensions of intact cells, as well as intact cell monolayers, glucosylate gelatinized collagen b) glucose transfer is not due to UDP-Glucose hydrolysis and subsequent intracellular utilization of the free glucose c) experiments using cell suspensions with known proportions of broken cells indicate that the glucosyltransferase activity is attributable to intact cells and not to contamination by intracellular enzymes from broken cells. The Km value for UDP-Glucose is about 20 microM. The enzyme has a pronounced requirement for manganese, and shows highest activity between 2 and 10 mM. The optimal Mn2+ concentration for the intracellular gelatin:glucosyltransferase activity is more restricted (5 to 10 mM). Glucosyltransferase activity is strongly inhibited by diamide and N-ethylmaleimide (5 mM), suggesting that intact sulfhydryl residues present in the enzyme are essential.     

Biosynthesis in vitro and the pattern of lectin binding receptors in monkey kidney cell surfaces.

The glycosyltransferase activities involved in the biosynthesis $\text{in vitro}$ of neutral blood group-related glycosphingolipids were measured in African green monkey kidney cells (Vero) grown in culture. The a-fucosyltransferases which catalyzed the reaction between GDP-fucose and corresponding acceptors to form H-active and novel Le^a-type glycosphingolipids were characterized in membrane fractions isolated from Vero cells and monkey bone marrow. Using ¹²⁵I-labeled $\text{Ulex europeus}$ and $\text{Lotus tetragonolobus}$ lectins the differential binding to Vero cell surface glycoproteins and glycolipids was studied under various conditions.     

Effects of brefeldin A on the localization of chondroitin sulfate-synthesizing enzymes. Activities in subfractions of the Golgi from chick embryo epiphyseal cartilage.

Membranes from brefeldin A-treated and untreated chick embryo epiphyseal cartilage were fractionated separately by equilibrium sucrose density gradient centrifugation. Fractions were assayed for Gal I transferase, Gal II transferase, Gal ovalbumin transferase, chondroitin polymerization on endogenous acceptors, GalNAc transfer to exogenous chondroitin hexasaccharide, and sulfate transfer to exogenous chondroitin. Gal I transferase and Gal II transferase activities were found in heavier cis- and medial-Golgi fractions, but with distributions different from each other. Brefeldin A had no effect on either their distribution or their total activity. Gal ovalbumin transferase activity in fractions from untreated cartilage was found as a dual peak in medial- and trans-Golgi areas. The latter peak was diminished in the fractions from the brefeldin A-treated cartilage, whereas the former peak was correspondingly increased. A similar dual medial- and trans-Golgi distribution for chondroitin polymerization on endogenous acceptors was seen with fractions from untreated cartilage. This was modified in fractions from brefeldin A-treated cartilage with a complete loss of synthesis in the trans-Golgi peak and a slight increase in synthesis in the medial-Golgi peak. However, the distribution of GalNAc transferase activity using exogenous chondroitin hexasaccharide indicated that considerable chondroitin-synthesizing activity still remained in these trans-Golgi fractions. This demonstrated that brefeldin A had caused a block in movement of endogenous proteochondroitin acceptors to the trans-Golgi site of synthesis. Sulfotransferase activity was also found in a dual distribution similar to that of the chondroitin polymerization and GalNAc transferase, with a small reduction in activity in the trans-Golgi fractions of brefeldin A-treated cartilage. Thus, treatment of cartilage with brefeldin A resulted in the loss of considerable trans-Golgi chondroitin sulfate-synthesizing enzyme activity and a block in the transport of one form of proteochondroitin precursor to the trans-Golgi membranes.