Wijnholt Ferwerda

Glycosylation of three molecular forms of human α1-acid glycoprotein having different interactions with concanavalin A

Human alpha 1-acid glycoprotein (AGP) was separated into a non-bound (AGP-A; 46%), a retarded (AGP-B; 39%) and a bound fraction (AGP-C; 15%) using concanavalin A (ConA)-Sepharose chromatography. The apparent molecular masses, as determined by SDS-PAGE, of the three fractions were 43.5, 42.3 and 41.2 kDa, respectively. The occurrence of N-linked di-, tri- and tetraantennary glycans on these three molecular forms (AGP-A, -B, and -C) was studied by sequential lectin-affinity chromatography of the 14C-labelled glycopeptides. These were obtained by extensive pronase treatment followed by N-[14C]acetylation of the peptide moieties. The glycopeptides of AGP-A did not bind to ConA-Sepharose whereas for AGP-B and AGP-C 18% and 44%, respectively, of the glycopeptides were bound as diantennary structures. Glycopeptide fractions of all three forms of AGP which were not bound to ConA-Sepharose were shown to contain equal amounts of both tri- and tetraantennary glycans by chromatography with Phaseolus vulgaris leukoagglutinating lectin (L-PHA). With the assumption that each molecule contains five glycosylation sites, it could be shown that AGP-A contains no diantennary structures whereas AGP-B and AGP-C contain one and two diantennary structures, respectively. In addition each of the molecular forms contains equal amounts of tri- and tetraantennary structures on the remaining glycosylation sites. The results of this study, therefore, exclude a uniformity of glycan chains in the three molecular forms of AGP. The degree of sialylation of each of the molecular forms was investigated by chromatography on L-PHA-agarose and Ricinus communis agglutinin-I--agarose both before and after desialylation of the glycopeptides. It was shown that about 90% of the biantennary glycans of both AGP-B and AGP-C were disialylated while the remainder were monosialylated. The degree of sialylation of the tri- and tetraantennary glycans was identical for the three molecular forms. In each case, one or more terminal galactose residues occurred on at least 20% of the tri- and 65% of the tetraantennary chains. It is suggested that the decrease in the exposure of galactose residues from AGP-A to AGP-C is related to the concomittant decrease in branching of the glycans of the three molecular forms. The relevance of these findings to studies on the function of AGP during inflammatory and liver diseases is discussed.

Turnover of Free Sialic Acid, CMP‐Sialic Acid, and Bound Sialic Acid In Rat Brain

Abstract: Adult male rats were injected intraventricularly with N‐[3H]acetylmannosamine. After different time intervals the rats were killed and free sialic acid, CMP‐sialic acid, lipid‐ and protein‐bound sialic acid were isolated from brain and the specific radioactivities determined. Maximal specific radioactivity was reached after approximately 4 h for CMP‐sialic acid, after 10–12 h for free sialic acid and after approximately 42 h for lipid‐and protein‐bound sialic acid. After some days the specific radioactivities of all four pools were the same and decreased equally, with a calculated turnover rate of approximately 3.5 weeks. The conclusion was that this phenomenon was the result of reutilisation of sialic acid and/or precursors. Therefore, the calculated turnover is not the turnover of bound sialic acid, but merely the rate of leakage of sialic acid and/or precursors out of the brain, so that no real turnover can be measured by this method. The first few hours after injection the specific radioactivity of CMP‐sialic acid rose above that of free sialic acid. It is supposed that a compartmentalization exists of free sialic acid. The newly synthesized sialic acid molecules are not secreted into the cytoplasmic pool but are preferentially used for the synthesis of CMP‐sialic acid. The results and conclusions are discussed in view of the general problems concerning turnover measurements of glycoconjugates.

Do antimetabolites interfere with the glycosylation of cellular glycoconjugates

INTRODUCTION ANTIMETABOLITES are frequently used for the treatment of various neoplastic diseases, including solid tumors and leukemias [l], but the use of these compounds is limited by the occurrence of myeloid and gastrointestinal toxicity. For most of these drugs the mechanism of action has been investigated extensively in vitro and in viva [l-5]. These studies, in general, point to interference in purine or pyrimidine biosynthesis (cf. Table 1) leading to inhibition of DNA and/or RNA synthesis in dividing cells. Additional mechanisms, however, might be responsible for the cytotoxic effects on non-dividing cells, i.e. the majority of normal cells and tumor cells which are in resting phase. Several antimetabolites interfere with the manufacture of nucleotides (UTP, GTP and CTP) that are substrates for the manufacture of nucleotide sugars. Since nucleotide sugars are the precursors for the biosynthesis of glycoconjugates (cf. Fig. l), the glycosylation of both secreted and membrane-associated glycoconjugates might be affected in the presence of antimetabolites. The sugar chains extending from the cell surface are important determinants of the phenotype of cells [6] and have been implicated in a variety of cell surface interactions [6-lo]. Characteristic changes in cell surface oligosaccharides accompany normal and malignant development and hematopoietic differentiation [ll-221. Thus, for a variety of malignant cells a characteristic increase in molecular weight, caused by changes in branching [ 1 l-171 and/or increased amount of sialic acid [ 11, 12, 16-181, has been observed for cell surface oligosaccharide structures in comparison with those from non-malignant cells. The invasive capacity of cells [17, 19, 201 and the loss of binding of malignant cells to-the substratum [12, 211 is related positively to the observed increase in molecular weight. In the last decade the evidence is compelling that the variations in oligosaccharide structures observed coincide with changes in activities of specific glycosyltransferases [e.g. 14, 21, 23-261, indicating the importance of these enzymes in the regulation of the glycosylation process. The general occurrence of characteristic changes in glycoconjugates in relation to malignancy and differentiation gives sup-

Subcellular and regional distribution of CMP-N-acetylneuraminic acid synthetase in the calf kidney

Abstract The subcellular localization of the enzyme CMP -N- acetylneuraminic acid (CMP-NANA) synthetase was studied in the different regions of the calf kidney. The enzyme appeared to be localized in the nuclear fractions. The possibility that CMP-NANA synthetase was adsorbed to the nuclear membrane during homogenization and subsequent isolation of the nuclei was excluded by removal of the membranes from purified nuclei by solubilization with Triton X-100. This treatment did not remove the CMP-NANA synthetase activity from the nuclei. About 11% of the CMP-NANA synthetase activity was recovered in the soluble fractions of the different regions of the calf kidney. To investigate whether this enzyme activity could be of nuclear origin too, some properties of the nuclear-bound and the soluble enzyme of the cortex of the kidney were compared. Both enzyme preparations exhibited identical pH and temperature optima. The apparent K m values for NANA, CTP and Mg 2+ for both enzyme preparations were also almost identical. Because (a) no differences were detectable between the two enzyme preparations and (b) it was observed that nuclei appear to be sensitive to homogenization, it is concluded that CMP-NANA synthetase recovered in the soluble fraction may be released from maltreated nuclei during homogenization.

Preparation and chemical characterization of calf tann-horsfall glycoprotein

Tamm-Horsfall glycoprotein preparations were obtained from calf urine by 1.0 M NaCl precipitation followed by 4 M urea/Sepharose 4B chromatography. By using 0.1% sodium dodecyl sulfate polyacrylamide gel electrophoresis a molecular weight of 86 500 +/- 4500 (n = 12) was calculated for the glycoprotein. Amino acid and carbohydrate analyses were performed, the carbohydrate composition being (in residues per 100 amino acid residues in the glycoprotein): fucose, 0.90; galactose, 4.82; mannose, 4.63;N-acetylglucosamine, 7.36; N-acetylgalactosamine, 1.38; sialic acid, 2.93. Under conditions of mild acid hydrolysis (0.05 M H2SO4, 80 degrees C, 1 h) the calf Tamm-Horsfall glycoprotein preparations were degraded partially into two lower molecular weight fragments (approximate Mr 66 000 and 51 000), as shown by polyacrylamide gel electrophoresis, both fragments being periodic acid-Schiff reagent positive.

INCORPORATION OF d‐[3H]GLUCOSAMINE INTO THE SIALOGLYCOPROTEIN GP‐350 OF ADULT RAT BRAIN

—The incorporation of d‐[3H]glucosamine into the nervous specific sialoglycoprotein GP‐350 has been studied in adult rat brain. Both the 100,000 g supernatant fluid and the 12,500 g pellet were used for the investigation, since GP‐350 could only be detected in the soluble cell fraction (Van Nieuw Amerongenet al., 1972) and in the synaptosomal membranes, sedimenting in the crude mitochondrial fraction (Van Nieuw Amerongen & Roukema, 1973, 1974). GP‐350 was separated from the other proteins by polyacrylamide gel electrophoresis at pH 7.5 and the incorporation of radioactivity into GP‐350 was measured at different time intervals, ranging from 1 to 96 h after the administration of the radioisotope. The maximal incorporation of radioactivity into the soluble GP‐350 was obtained after about 2 h and into the membrane‐bound GP‐350 after about 3 h. After 2 h there is a very rapid decrease of the radioactivity of GP‐350 from the soluble cell fraction (up to 70 per cent within 2 h). Thereafter the disappearance is more gradual and of the same order as that found for the membrane‐bound fraction of GP‐350. The half‐life of the soluble GP‐350 was estimated to be 19 h and for the membrane‐bound GP‐350 a value of 18 h was calculated. Compared to the total pool of brain (glyco) proteins and specific nervous (glyco) proteins GP‐350 has a very rapid turnover. The rapid initial decrease of the radioactivity from the soluble GP‐350 may be interpreted in terms of a rapid transport of the newly‐synthesized GP‐350 from the cytoplasma of the perikaryon to the membranes of the synaptic region by an axoplasmic flow.

CMP-N-acetylneuraminic acid: Is it synthesized in the nucleus

RadioactiveN-acetylmannosamine was injected intravenously into rats to labelN-acetylneuraminic acid (NeuAc) and CMP-NeuAc. Nuclei were isolated from the livers using a non-aqueous technique to prevent leakage of polar metabolites. A preparation was obtained, which was eight times enriched in nuclei based on the ratio DNA/RNA. Free NeuAc and CMP-NeuAc were isolated from this nuclear fraction and from whole liver, and the specific radioactivities were determined. It appeared that at all time points studied, i.e. 1.5, 9.5, and 18 min after injection, the specific radioactivities of free NeuAc as well as of CMP-NeuAc in the nuclear preparation were lower than those in whole liver. Also no significant differences were found between free NeuAc and CMP-NeuAc in the ratio of specific radioactivities in the nuclear fraction/whole liver. Furthermore, no enzyme involved in the synthesis of NeuAc was enriched in the nuclear preparation as compared to various other cytosolic and non-cytosolic enzymes.Because newly synthesized NeuAc is channelled into a special pool and used for activation to CMP-NeuAc [Ferwerda W, Blok CM, van Rinsum J (1983) Biochem J 216:87–92], these results point to a site of activation of NeuAc to CMP-NeuAc other than the nuclear compartment. This might indicate that the nuclear-localized enzyme, CMP-NeuAc synthase, leaves the nucleus before exerting its action.

N-Acetylneuraminic acid biosynthesis in rat liver and kidney

Rat liver and kidney tissue slices incubated withN-acetyl [3H]mannosamine incorporated radioactivity into free and boundN-acetylneuraminic acid and CMP-N-acetylneuraminic acid (CMP-NeuAc). Liver and kidney also incorporated radioactivity from intravenously injected [3H]ManNAc intoN-acetylneuraminic acid and CMP-NeuAc. From the decrease in the specific radioactivity of CMP-NeuAc after a single injection ofN-acetyl[3H]mannosamine the half-life of CMP-NeuAc was determined. From this half-life and the pool size of CMP-NeuAc a synthesis rate of CMP-NeuAc was calculated, being 1.2 nmol/min/g wet weight of kidney. In previous experiments a value of 1.0 nmol/min/g wet weight was determined for liver [Ferwerdaet al. (1983) Biochem J 216: 87–92]. The synthesis rate of CMP-NeuAcin vivo was in the same range as the synthesis rate calculated from the turnover of boundN-acetylneuraminic acid, which was 2.7 and 0.4 nmol/min/g wet weight for liver and kidney respectively.The assay conditions for UDP-N-acetylglucosamine 2-epimerase andN-acetylmannosamine kinase were adapted to measure low activitiesin vitro. It appeared that the kinase activity detected in kidney can synthesizeN-acetylmannosamine6-phosphate at a rate sufficient for the observed production ofN-acetylneuraminic acidin vivo. Also a low, but measurable activity of UDP-N-acetylglucosamine 2-epimerase was detected in kidneyin vitro, suggesting that the biosynthetic pathway ofN-acetylneuraminic acid in kidney is the same as in liver. The synthesis rate ofN-acetylneuraminic acid in liver determinedin vivo is approximately 12 times slower than the maximal potential rate calculated from the activities of theN-acetylneuraminic acid (precursor-) forming enzymes as detectedin vitro. This indicates that in liverin vivo the enzymes are working far below their maximal capacity.

125 INCORPORATION OF 5-FLUOROURACIL INTO NUCLEOTIDE SUGARS AND THE EFFECT ON GLYCOCONJUGATES IN RAT HEPATOMA CELLS AND HEPATOCYTES

Liver is the main organ in the elimination of 5-fluorouracil (5FU). Rat hepatocytes and hepatoma cells (H35) were incubated in monolayer in the presence of 0, 17, 50 and 100 μM 5FU for 4 hrs. The metabolism of 5FU was monitored with [3H]5FU, the protein synthesis was determined by the incorporation of [14C]leucine and glycoconjugate synthesis by incorporation of [3H]GlcNH2 and [3H]fucose. In rat hepatocytes the amount of soluble fluoronu-cleotides was dependent on the concentration of 5FU. With 50 μM 5FU the concentrations of FUDP-HexNAc, FUDP-hexose and FUTP were 17, 5 and 9 pmol/106 hepatocytes. In the presence of 0.5 mM thymine to inhibit the catabolism of 5FU, these concentrations were 23, 29 and 49 pmol/106 hepatocytes, resp. The effect of thymine at 17 and 100 μM 5FU was comparable. The incorporation of 5FU into RNA was also concentration dependent and was 72 pmol/106 hepatocytes at 50 μM 5FU. In contrast to nucleotide synthesis thymine enhanced the incorporation of 5FU into RNA only less than 1.5 fold at all 5FU concentrations. In H35 cells both synthesis of soluble fluoronucleotides and the incorporation of 5FU into RNA was linear with the concentration of 5FU. At 50 μM 5FU the amounts of FUDP-HexNAC, FUDP-hexose and FUTP were 30, 68 and 106 pmol/106 cells, and 345 pmol 5FU/106 cells was incorporated into RNA. In both cell types incubation with 5FU had no effect on the incorporation of leucine into proteins and of sugars into glycoconjugates.