Portal de Periódicos da CAPES

3′-Azidothymidine Potently Inhibits the Biosynthesis of Highly Branched N-Linked Oligosaccharides and Poly-N-acetyllactosamine Chains in Cells

2000; Elsevier BV; Volume: 275; Issue: 35 Linguagem: Inglês

10.1016/s0021-9258(19)61448-1

1083-351X

Richard Steet, Paul Melançon, Robert D. Kuchta,

Proteoglycans and glycosaminoglycans research

Previous studies in our laboratory have characterized 3′-azido-3′-deoxythymidine (AZT) as a potent inhibitor of glycosphingolipid biosynthesis in cultured cells (Steet, R., Alizadeh, M., Melançon, P., and Kuchta, R. D. (1999) Glycoconj. J. 16, 237–245; Yan, J.-P., Ilsley, D. D., Frohlick, C., Steet, R., Hall, E. T., Kuchta, R. D., and Melançon, P. (1995) J. Biol. Chem. 270, 22836–22841). Here, we report that AZT treatment of K562 cells results in significant alterations in the profile of N-linked oligosaccharides. Fractionation of [3H]mannose-labeled oligosaccharides from AZT-treated K562 cells using lectin affinity chromatography revealed striking changes in the branching and processing ofN-linked glycoconjugates. AZT treatment resulted in the production of fewer highly branched complex glycans (60% of control at 20 μm AZT) and a significant accumulation of core-fucosylated biantennary oligosaccharides. In addition, extension of branched oligosaccharides with multiple poly-N-acetyllactosamine repeats is nearly abolished by AZT concentrations as low as 2 μm. A shift from multiantennary to moderately branched oligosaccharides was also apparent in the melanoma cell line SK-MEL-30 upon AZT treatment.N-Linked glycans from both cell lines exhibited increased affinity for the β-galactoside-binding lectin RCA-I in the presence of AZT, suggesting that the addition of terminal sialic acid is sensitive to the drug. These results demonstrate the ability of AZT to modulate strongly the processing of asparagine-linked glycoconjugates in whole cells and reveal a novel mechanism by which AZT treatment may cause anemia. Previous studies in our laboratory have characterized 3′-azido-3′-deoxythymidine (AZT) as a potent inhibitor of glycosphingolipid biosynthesis in cultured cells (Steet, R., Alizadeh, M., Melançon, P., and Kuchta, R. D. (1999) Glycoconj. J. 16, 237–245; Yan, J.-P., Ilsley, D. D., Frohlick, C., Steet, R., Hall, E. T., Kuchta, R. D., and Melançon, P. (1995) J. Biol. Chem. 270, 22836–22841). Here, we report that AZT treatment of K562 cells results in significant alterations in the profile of N-linked oligosaccharides. Fractionation of [3H]mannose-labeled oligosaccharides from AZT-treated K562 cells using lectin affinity chromatography revealed striking changes in the branching and processing ofN-linked glycoconjugates. AZT treatment resulted in the production of fewer highly branched complex glycans (60% of control at 20 μm AZT) and a significant accumulation of core-fucosylated biantennary oligosaccharides. In addition, extension of branched oligosaccharides with multiple poly-N-acetyllactosamine repeats is nearly abolished by AZT concentrations as low as 2 μm. A shift from multiantennary to moderately branched oligosaccharides was also apparent in the melanoma cell line SK-MEL-30 upon AZT treatment.N-Linked glycans from both cell lines exhibited increased affinity for the β-galactoside-binding lectin RCA-I in the presence of AZT, suggesting that the addition of terminal sialic acid is sensitive to the drug. These results demonstrate the ability of AZT to modulate strongly the processing of asparagine-linked glycoconjugates in whole cells and reveal a novel mechanism by which AZT treatment may cause anemia. 3′-azido-3′-deoxythymidine 3′-azido-3′-deoxy-thymidine monophosphate concanavalin A Datura stramoniumagarose N-acetylglucosaminyltransferase Griffonia simiplicifolia-II lectin N-acetyllactosamine leuko-phytohemagglutinin mannose phosphate-buffered saline Pisum sativum(pea) agarose Ricinus communis agglutinin-1 Recent studies in our laboratory (1Steet R. Alizadeh M. Melançon P. Kuchta R.D. Glycoconj. J. 1999; 16: 237-245Crossref PubMed Scopus (10) Google Scholar, 2Yan J.-P. Ilsley D.D. Frohlick C. Steet R. Hall E.T. Kuchta R.D. Melançon P. J. Biol. Chem. 1995; 270: 22836-22841Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar) have established the thymidine analog 3′-azidothymidine (AZT)1 as a potent inhibitor of lipid and protein glycosylation reactions in whole cells. This inhibition likely results from the ability of AZTMP, the primary intracellular metabolite of AZT, to inhibit competitively pyrimidine nucleotide-sugar import into the Golgi lumen and, consequently, modify the macromolecular glycosylation reactions that occur in this organelle (3Hall E.T. Yan J.-P. Melançon P. Kuchta R.D. J. Biol. Chem. 1994; 269: 14355-14358Abstract Full Text PDF PubMed Google Scholar). AZT treatment of the erythroleukemia cell line K562 and a variety of melanoma cell lines resulted in dramatic alterations in glycosphingolipid biosynthesis, namely the severe inhibition of ganglioside synthesis (1Steet R. Alizadeh M. Melançon P. Kuchta R.D. Glycoconj. J. 1999; 16: 237-245Crossref PubMed Scopus (10) Google Scholar, 2Yan J.-P. Ilsley D.D. Frohlick C. Steet R. Hall E.T. Kuchta R.D. Melançon P. J. Biol. Chem. 1995; 270: 22836-22841Abstract Full Text Full Text PDF PubMed Scopus (56) Google Scholar). Indeed, significant decreases in complex ganglioside synthesis were apparent at AZT concentrations as low as 1 μm. Furthermore, AZT treatment of K562 cells reduced the incorporation of galactose and sialic acid into proteins. Some of the side effects associated with AZT treatment likely result from the ability of the drug to block the early development of blood progenitor stem cells (anemia and neutropenia (4Richman D.D. Fischl M.A. Grieco M.H. Gottlieb M.S. Volberding P.A. Laskin O.L. Leedom J.M. Groopman J.E. Mildvan D. Hirsch M.S. Jackson G.G. Durack D.T. Nusinoff-Lehrman S. N. Engl. J. Med. 1987; 317: 192-197Crossref PubMed Google Scholar, 5Dainiak N. Worthington M. Riordan M.A. Kreczko S. Goldman L. Br. J. Haematol. 1988; 69: 299-304Crossref PubMed Scopus (101) Google Scholar, 6Du D.L. Volpe D.A. Grieshaber C.K. Murphy M.J.J. Br. J. Haematol. 1992; 80: 437-445Crossref PubMed Scopus (26) Google Scholar, 7Ganser A. Greher J. Völkers B. Staszewski S. Hoelzer D. Exp. Hematol. 1989; 17: 321-325PubMed Google Scholar)). An integral component of the differentiation program of many eukaryotic cells, changes in glycosylation are strongly associated with the development of hematopoietic stem cells. For example, blood progenitor cell differentiation is characterized by pronounced stage-specific modifications in the structure of the asparagine-linked oligosaccharides of several cell surface proteins (8West C.M. Mol. Cell. Biochem. 1986; 72: 3-20Crossref PubMed Scopus (100) Google Scholar). Several other developmentally regulated events including increases inO-linked glycosylation, changes in proteoglycan expression, and altered glycosphingolipid metabolism also occur concomitant with the maturation of many hematopoietic stem cells (9Gahmberg C.G. Ekblom M. Andersson L.C. Proc. Natl. Acad. Sci. U. S. A. 1984; 81: 6752-6756Crossref PubMed Scopus (51) Google Scholar, 10Gahmberg C.G. Autero M. Hermonen J. J. Cell. Biochem. 1988; 37: 91-105Crossref PubMed Scopus (18) Google Scholar, 11Drzeniek Z. Stöcker G. Siebertz B. Just U. Schroeder T. Ostertag W. Haubeck H.D. Blood. 1999; 93: 2884-2897Crossref PubMed Google Scholar, 12Hakomori S.-I. Annu. Rev. Biochem. 1981; 50: 733-764Crossref PubMed Scopus (1473) Google Scholar). Although the precise role of these changes has not been completely elucidated, the remodeling of surface oligosaccharides by inhibitors of glycosylation may compromise the ability of these cells to proliferate and differentiate by altering specific structural determinants crucial for cell-cell interactions and signal transduction events. Clinical and biochemical studies have provided clear evidence that altering N-linked glycosylation alone can lead to the failure of blood stem cells to mature properly. Inherited anemias, such as congenital dyserythropoietic anemia type II or hereditary erythroblastic multinuclearity with positive acidified serum lysis test, are associated with reduced activity of GnT-II or α-mannosidase II (13Fukuda M.N. Dell A. Scartezzini P. J. Biol. Chem. 1987; 262: 7195-7206Abstract Full Text PDF PubMed Google Scholar, 14Fukuda M.N. Masri K.A. Dell A. Luzzatto L. Moremen K.W. Proc. Natl. Acad. Sci. U. S. A. 1990; 87: 7443-7447Crossref PubMed Scopus (94) Google Scholar), enzymes involved in the processing of complexN-glycans. Erythrocyte proteins from afflicted individuals lacked poly-LacNAc additions, and as a result, those cells exhibited overall membrane instability (15Fukuda M.N. Papayannopoulou T. Gordon-Smith E.C. Rochant H. Testa U. Br. J. Haematol. 1984; 56: 55-68Crossref PubMed Scopus (58) Google Scholar, 16Wickramasinghe S.N. Blood Rev. 1998; 12: 178-200Crossref PubMed Scopus (79) Google Scholar, 17Fukuda M.N. Klier G., Yu, J. Scartezzini P. Blood. 1986; 68: 521-529Crossref PubMed Google Scholar). More recently, genetic disruption of the α-mannosidase II gene in a mouse model was shown to result in dyserythropoiesis (18Chui D. Oh-Eda M. Liao Y.-F. Penneerselvam K. Lai A. Marek K.W. Freeze H.H. Moreman K.W. Fukuda M.N. Marth J.D. Cell. 1997; 90: 157-167Abstract Full Text Full Text PDF PubMed Scopus (183) Google Scholar), thereby providing further evidence of the essential role of complex N-glycans in the development of pluripotent blood stem cells. Significant alterations in N-linked oligosaccharide structure are also evident during oncogenic transformation (19Hakomori S. Cancer. Res. 1996; 56: 5309-5318PubMed Google Scholar). These changes are closely associated with tumor progression and often correlate with the metastatic behavior of many malignancies (20Dennis J.W. Fukuda M. Cell Surface Carbohydrates and Cell Development. CRC Press, Inc., Boca Raton, FL1992: 161-194Google Scholar, 21Asada M. Furukawa K. Segawa K. Endo T. Kobata A. Cancer Res. 1997; 57: 1073-1080PubMed Google Scholar). For example, an elevation in multiantennary complex glycans containing GlcNAc-(β1,6)-Man branches is often accompanied by increased metastatic capacity (22Dennis J.W. LaFerte S. Waghorne C. Breitman M.L. Kerbel R.S. Science. 1987; 236: 582-585Crossref PubMed Scopus (857) Google Scholar). In addition, several studies have demonstrated that enhanced sialylation of highly branchedN-linked oligosaccharides may promote tumor progression and metastasis (23Takano R. Muchmore E. Dennis J.W. Glycobiology. 1994; 4: 665-674Crossref PubMed Scopus (75) Google Scholar, 24Yogeeswaran G. Salk P.L. Science. 1981; 212: 1514-1516Crossref PubMed Scopus (408) Google Scholar, 25Bresalier R.S. Rockwell R.W. Dahiya R. Duh Q.Y. Kim Y.S. Cancer Res. 1990; 50: 1299-1307PubMed Google Scholar). In light of the importance of protein glycosylation for both erythrocyte maturation and tumor biology, we examined the effects of AZT on N-linked glycosylation in two cell lines, the erythroleukemia cell line K562 and the melanoma cell line SK-MEL-30. Lectin affinity analysis of radiolabeled glycopeptides from AZT-treated cells revealed a profound decrease in the amount of highly branched complex N-glycans, severe suppression of long chain poly-LacNAc synthesis, and a reduction in terminal sialylation. The ability of AZT to alter the branching and processing ofN-linked oligosaccharides and the potential for these changes to influence the metastatic behavior of tumor cells as well as the relevance of these findings to the hematological toxicity accompanying AZT therapy are discussed. K562 and SK-MEL-30 cells were obtained from the American Type Culture Collection and maintained in RPMI 1640 supplemented with 10% fetal calf serum, 100 mg/liter penicillin G, and 100 mg/liter streptomycin at 37 °C in a humidified atmosphere of 95% air, 5% CO2. Cell culture media and bovine fetal serum were obtained from Life Technologies, Inc.d-[2-3H]Mannose (10–20 Ci/mmol), Tran35S metabolic label (1175 Ci/mmol), and [14C]mannose were purchased from American Radiolabeled Chemicals. ConA-Sepharose was from Amersham Pharmacia Biotech, and all other agarose-bound lectins employed in this study were from Vector Laboratories. Bio-Gel P-4 was purchased from Bio-Rad. QAE-Sephadex, alkaline phosphatase, β-galactosidase, Arthrobacter ureafaciens neuraminidase, α-methylmannoside, α-methylglucoside, d-mannose, and l-fucose were obtained from Sigma, and AZT was from ICN. Pronase andN-glycosidase F were purchased from Roche Molecular Biochemicals. Escherichia freundiiendo-β-galactosidase was a generous gift from Dr. Michiko Fukuda, La Jolla Cancer Research Foundation. K562 cells (5.0 × 105 cells/well in 1.5 ml of media) were labeled with either 1.0 μCi of Tran35S metabolic label or 6 μCi of [3H]mannose and grown for 24 h at 37 °C. The cells were collected on ice and washed four times with PBS before either being counted directly or collected on glass filters. The collected cells were washed seven times with 5% trichloroacetic acid and dried, and radioactivity was measured by liquid scintillation counting. In some cases, cells were collected, washed with PBS, and extracted with organic solvents (see below) to remove lipid-linked species. The extracted material was dried under a stream of N2 (g), solubilized in scintillation fluid, and radioactivity determined by scintillation counting. The remaining protein pellet was washed with water, solubilized in 1% SDS, and radioactivity determined as before. The CHCl3/MeOH (2:1) extracts from [3H]mannose-labeled K562 cells were pooled and resolved on silica TLC plates developed with CHCl3/MeOH/H2O (60:35:8). The plates were carefully cut into 1-cm strips and solubilized in scintillation fluid for several hours prior to counting. For analysis of the second organic extraction, the procedure of Suzuki and Lennarz (26Suzuki T. Lennarz W.J. Glycobiology. 2000; 10: 51-58Crossref PubMed Scopus (13) Google Scholar) was used with only slight modifications. Briefly, the CHCl3/MeOH/H2O (10:10:3) extracts from [3H]mannose-labeled K562 cells were pooled and dried by rotary evaporation under reduced pressure. The samples were resuspended in 80% tetrahydrofuran, 0.1 n HCl and heated at 65 °C for 1 h to release the oligosaccharides. Following neutralization with NaOH, the samples were dried, resuspended in 10 mmTris, pH 8.0, and applied to a column of Bio-Gel P4 (1.0 × 100 cm). Elution was carried out in water, and fractions (1.0 ml) were collected, and radioactivity was monitored by liquid scintillation counting. The column was calibrated using blue dextran and [14C]mannose. The concentration and specific activity of GDP-mannose in K562 cells were obtained using the method of Rush and Waechter (27Rush J.S. Waechter C.J. Anal. Biochem. 1995; 224: 494-501Crossref PubMed Scopus (16) Google Scholar). Briefly, 20 × 106 cells in 24 ml of RPMI 1640 were metabolically labeled for 4 h with [3H]mannose (10 μCi/ml). To determine the effects of AZT on GDP-mannose, cells were treated with 20 μm AZT prior to labeling. The labeled cells were collected, washed with PBS (3 × 5 ml), and extracted twice with 0.3 ml of 70% EtoH containing 10 mm ammonium phosphate, pH 3.5. Cellular GDP-mannose was partially purified from the cell extracts by chromatography on ConA-Sepharose. Retained fractions containing labeled GDP-mannose were pooled and further purified by anion exchange chromatography on a Partisil-10 SAX column. GDP-mannose concentrations were determined by calculation of the area under theA 254 peak corresponding to GDP-mannose and comparison with known amounts. K562 cells were labeled by growing 5.5–6 × 106 cells in the presence of 15–25 μCi/mld-[2-3H]mannose for 24 h in 12 ml of RPMI 1640. SK-MEL-30 cells (3.0 × 106) were labeled with d-[2-3H]mannose (10 μCi/ml) for 30 h in 12 ml of RPMI 1640. In both cases, 0–20 μmAZT were present throughout the entire labeling period. The preparation of glycopeptides was done essentially as described by Fujimoto and Kornfeld (28Fujimoto K. Kornfeld R. J. Biol. Chem. 1991; 266: 3571-3578Abstract Full Text PDF PubMed Google Scholar). Briefly, cells were collected by centrifugation at 120 × g for 5 min at 5 °C. The resulting cell pellets were washed twice with ice-cold PBS and then extracted as follows: 3 × 3 ml of CHCl3/MeOH (2:1) with sonication, 2 × 3 ml of H2O, and three times sequentially with 3, 2, and 2 ml of CHCl3/MeOH/H2O (10:10:3), each time pelleting the residue by centrifugation for 15 min (1000 × g). The pellets were dried under a stream of N2 between different solvent extractions. For analysis of lipids, the extracts were pooled for further evaluation as described above. The extracted cell pellets were next digested with Pronase (2 mg/ml) in 2 ml of 100 mm Tris, pH 8.0, 20 mm CaCl2 at 57 °C for 60 h to generate glycopeptides. The Pronase digestions were stopped by boiling the samples for 3 min, and the glycopeptides were desalted on a column of Sephadex G-25 (1.0 × 50 cm) equilibrated in H2O prior to lectin affinity analysis. All lectin chromatography was performed at 4 °C with flow rates determined by gravity. The desalted radiolabeled glycopeptides were first fractionated on ConA-Sepharose (2 ml bed) equilibrated in 10 mm Tris, pH 8.0, 150 mm NaCl, 1 mmCaCl2, and 1 mm MnCl2 (buffer A). Samples in this buffer were loaded onto the columns, washed with buffer A, and then eluted with 10 mm α-methylglucoside in buffer A followed by 200 mm α-methylmannoside in buffer A prewarmed to 55 °C. An aliquot of each fraction was mixed with scintillation fluid and analyzed for radioactivity. Fractions containing radioactivity were then pooled and designated as ConA pools I, II, III. The glycopeptides from ConA pool I were desalted on Sephadex G-25, dried, taken up in 10 mm HEPES, pH 7.5, 150 mm NaCl, and 0.1 mm CaCl2 (buffer B), and applied to columns of agarose-bound tomato lectin (0.7 × 15 cm) equilibrated in buffer B. Columns were eluted with undiluted chitin hydrolysate (Vector Laboratories) and fractions monitored by liquid scintillation counting. ConA pool I glycopeptides were also applied to columns of immobilized L-PHA (0.7 × 20 cm) equilibrated in buffer B containing 1 mm MnCl2(buffer C) and eluted with buffer C. Fractionation of ConA pools I and II using agarose-bound pea lectin was performed on columns (1.0 × 5.0 cm) equilibrated in buffer C. Columns were eluted with 200 mm α-methylmannoside, 200 mmα-methylglucoside in buffer C. Analysis of ConA pools I and II using GSL II and RCA-immobilized lectins was performed by applying samples to columns (0.7 × 15 cm) equilibrated in HEPES-buffered saline, pH 7.5 (buffer D). RCA-agarose-bound glycopeptides were eluted with 200 mmlactose in buffer D. Chitin hydrolysate diluted 2-fold in buffer D was used to elute bound glycopeptides from immobilized GSL II. Collected fractions were monitored for radioactivity as above. Samples were desalted on a Sephadex G-25 column, pooled, and adjusted to 50 mm sodium phosphate, pH 7.2, and 5 mm EDTA. Samples were digested with 5 units ofN-glycosidase F overnight at 37 °C in order to liberate the charged peptide. Samples were boiled for 3 min to stop the reaction and desalted on a Sephadex G-25 column. In all cases, a shift in the elution of the radioactivity corresponding to an overall lower molecular weight was observed confirming the release of the peptide. Columns containing 1.5 ml of QAE-Sephadex were equilibrated with 2 mm Tris base, and samples brought up in 2–3 ml of this buffer were applied to the columns. Charged material was eluted from the column with increasing concentrations of NaCl dissolved in 2 mm Tris base. A portion of the peptide-liberated,3H-labeled oligosaccharides was dried by rotary evaporation and dissolved in 50 μl of 100 mm Tris-HCl, pH 8.0, and treated with 5 units of alkaline phosphatase from bovine mucosa for 1 h at 37 °C. Treated and untreated samples were taken up in 2 ml of 2 mm Tris base and chromatographed on QAE-Sephadex (see above). Samples (4,000–5,000 cpm) were treated with 0.05 units of enzyme in 0.5 ml of 20 mm potassium phosphate, pH 7.2, for 12 h at 37 °C. ConA pool I samples (1,000–1,500 cpm) were incubated with 25 milliunits of enzyme in 0.75 ml of 50 mm NaOAc, pH 5.7, and 250 μg/ml BSA for 20 h at 37 °C. Desialylated ConA pool I samples were incubated with 2 units of β-galactosidase in 1.0 ml of 100 mm NaOAc, pH 5.5, for 18 h at 37 °C. Dried portions (1,000–3,000 cpm) of ConA pool I glycopeptides were dissolved in 50 μl of 10 mm HCl and heated at 100 °C for 30 min to remove sialic acids. Samples of desalted glycopeptides from [3H]mannose-labeled K562 cells were hydrolyzed in 1 m HCl at 100 °C for 4 h. After neutralization, the samples were dried, brought up in 10 μl of H2O, and loaded onto Whatman No. 1 paper. Descending paper chromatography was performed using EtoAc/pyridine/HOAc/H2O (5:5:1:3), and sugar standards were visualized using alkaline silver nitrate (29Trevelyan W.E. Proctor D.P. Harrison J. Nature. 1950; 166: 444-445Crossref PubMed Scopus (2201) Google Scholar). Radiolabeled sugars were detected by cutting the paper into 1-cm strips and placing the pieces in vials containing 0.7 ml of H2O for 1 h followed by the addition of 3.5 ml of scintillation fluid for scintillation counting. Previous work in our laboratory has shown that AZT treatment of K562 cells inhibits galactose and sialic acid incorporation into trichloroacetic acid-precipitable material. The reduction in galactose and sialic acid incorporation could result from a loss or modification to either N- or O-linked surface oligosaccharides. In order to address whether cellularN-linked glycosylation was sensitive to AZT treatment, [3H]mannose-labeled oligosaccharides isolated from control and AZT-treated K562 cells were subjected to serial lectin analysis. Initially, the radiolabeled glycopeptides were fractionated into three pools using ConA-Sepharose. The glycopeptide fraction in ConA pool I, which does not bind the lectin, is expected to primarily contain tetra- and triantennary complex structures, whereas ConA pool II likely consists of biantennary complex glycans and unusual hybrid structures (30Cummings R.D. Kornfeld S. J. Biol. Chem. 1982; 257: 11235-11240Abstract Full Text PDF PubMed Google Scholar). ConA pool III will contain hybrid and oligomannose glycans that bind with high affinity to this lectin due to the presence of exposed mannose residues. Table I shows that AZT treatment results in a striking remodeling of the N-linked oligosaccharides synthesized by K562 cells. Most notably, the relative amount of highly branched complex structures (ConA pool I) was significantly reduced. With 20 μm AZT, the fraction of mannose found in ConA pool I decreased by 39%. In addition, AZT treatment led to an accumulation of both biantennary complex glycans (ConA pool II) as well as hybrid-type and oligomannose structures (ConA pool III). These effects of AZT on the degree of branching in complex glycans (ConA I and II) occurred in a dose-dependent manner and are apparent at concentrations of the drug as low as 5 μm.Table IDistribution of [ 3 H]mannose-labeled glycopeptides from K562 cells synthesized in the presence or absence of AZTLectinFraction0 μm AZT2 μmAZT5 μm AZT20 μm AZT% total cpm in K562 glycopeptide fractionsConA (24 h)I36 ± 335 ± 232 ± 322 ± 2II7 ± 19 ± 210 ± 214 ± 1III57 ± 556 ± 458 ± 564 ± 2PSAII-UB4.0 ± 0.54.2 ± 0.34.8 ± 0.45.7 ± 0.6II-B3.5 ± 0.24.5 ± 0.35.2 ± 0.28.3 ± 0.7L-PHAI-UB18 ± 118 ± 217 ± 111.9 ± 0.7I-B18 ± 117 ± 215 ± 210.0 ± 0.5ConA (6 h)I4125II 715III5260[3H]Mannose-labeled glycopeptides obtained from K562 cells treated with various AZT concentrations for 6 or 24 h were fractionated initially using ConA-Sepharose. As noted, portions of ConA pools I and II were further chromatographed using immobilized L-PHA and PSA lectins (UB is unbound; B is bound). For L-PHA and PSA, the sum of the "% cpm" will add up to % cpm in the ConA pool applied to each lectin. Data represent the mean from at least two independent experiments. Open table in a new tab [3H]Mannose-labeled glycopeptides obtained from K562 cells treated with various AZT concentrations for 6 or 24 h were fractionated initially using ConA-Sepharose. As noted, portions of ConA pools I and II were further chromatographed using immobilized L-PHA and PSA lectins (UB is unbound; B is bound). For L-PHA and PSA, the sum of the "% cpm" will add up to % cpm in the ConA pool applied to each lectin. Data represent the mean from at least two independent experiments. The ability of AZT to alter the relative distribution ofN-linked oligosaccharides in K562 cells after brief drug exposure was examined to test the possibility that the effects of AZT were due to alterations in the biosynthesis of the glycosylation machinery (glycosyltransferases, etc.), as opposed to direct effects on glycosylation. As shown in Table I, the ratios of the pools observed upon ConA chromatography of treated (20 μm AZT) and control samples were nearly identical at 6- and 24-h labeling times. Since the effects of AZT are apparent after brief exposure to the drug, it is likely that the observed alterations do not result from changes in the levels of enzymes involved in glycosylation. The effects of AZT on the ConA profile were not due to effects on the interconversion of mannose and fucose. Metabolic labeling of cells with [2-3H]mannose normally results in the radiolabel being incorporated into macromolecules as both mannose and fucose (31Varki A. Methods Enzymol. 1994; 230: 16-34Crossref PubMed Scopus (56) Google Scholar). Total cellular glycopeptides were generated from [3H]mannose-labeled cells and subjected to acid hydrolysis to convert all oligosaccharides into monosaccharides. Analysis of the hydrolysates by descending paper chromatography revealed that the fraction of radiolabel present as mannose from cells treated with either no AZT (0.75 ± 0.03) or 20 μmAZT (0.73 ± 0.02) was identical. AZT treatment resulted in a modest dose-dependent reduction in the total amount of mannose incorporated into proteins (Table II). Importantly, the magnitude of this reduction is relatively small compared with the extent of changes observed upon ConA fractionation. For example, at 20 μm AZT, total mannose incorporation declined by only 16%, whereas the fraction of mannose recovered in ConA pool I decreased by nearly 40%. The reduced incorporation was not due to an effect on polypeptide synthesis, since AZT concentrations as high as 20 μm had no effect on protein synthesis after 24 h of treatment as measured by 35S metabolic labeling of cells.Table IIEffect of AZT on the incorporation of [ 3 H]mannose and Tran 35 S metabolic label into lipid and protein-linked oligosaccharides from K562 cells[AZT]Protein-linkedLipid-linked [3H]Man[3H]Man[35S]Met/Cysμmtotal cpm recovered (% control)0100 ± 3100 ± 2100 ± 2296 ± 4ND2-aNot determined.100 ± 2591 ± 6ND103 ± 12084 ± 9104 ± 599 ± 1K562 cells were grown in RPMI 1640 media containing [3H]mannose or [35S]Met/Cys metabolic label in the presence or absence of various concentrations of AZT for 24 h. The percent of radioactivity in each sample relative to that contained in control cells is shown and represents the average of three separate experiments. The lipid-linked fraction represents the total cpm recovered from both extractions (see "Experimental Procedures").2-a Not determined. Open table in a new tab K562 cells were grown in RPMI 1640 media containing [3H]mannose or [35S]Met/Cys metabolic label in the presence or absence of various concentrations of AZT for 24 h. The percent of radioactivity in each sample relative to that contained in control cells is shown and represents the average of three separate experiments. The lipid-linked fraction represents the total cpm recovered from both extractions (see "Experimental Procedures"). We performed several control experiments to determine if the decreased incorporation of mannose involved changes in the synthesis of the dolichol-linked core oligosaccharide. Initially, we demonstrated that AZT did not affect GDP-[3H]mannose pools. K562 cells were labeled with [3H]mannose and the concentration and specific activity of GDP-[3H]mannose determined. The concentration and specific activity of GDP-mannose present in control cells (2.55 ± 0.20 pmol/106 cells, 34.7 ± 3.3 cpm/pmol) was nearly identical to cells treated with 20 μm AZT (2.45 ± 0.25 pmol/106 cells and 33.8 ± 2.8 cpm/pmol), indicating that the decreased incorporation of mannose into glycoproteins upon AZT treatment was not due to changes in GDP-[3H]mannose pools. To exclude the possibility that the reduction in mannose incorporation reflected changes in dolichol-sugar synthesis, the effects of AZT on the amount and type of lipid-linked oligosaccharide present in K562 cells was measured. In contrast to protein-associated label, there was no significant change in the total amount of [3H]mannose incorporated into lipid-linked molecules upon AZT treatment (Table II). Analysis of a 2:1 CHCl3/MeOH extract by silica TLC revealed that the majority (>75%) of the recovered radioactivity was present as single band, most likely dolichol-P-mannose (data not shown, (32Pan Y.T. Elbein A.D. Biochemistry. 1990; 29: 8077-8084Crossref PubMed Scopus (20) Google Scholar)). Importantly, treatment with 20 μm AZT did not affect either the amount or mobility of this species. Furthermore, analysis by gel filtration of the second extract (10:10:3 CHCl3/MeOH/H2O), enriched in dolichol-linked oligosaccharides (33Rosenwald A.G. Stoll J. Krag S.S. J. Biol. Chem. 1990; 265: 14544-14553Abstract Full Text PDF PubMed Google Scholar), also showed minimal differences between treated and untreated cells. In both cases (Fig.1), the majority (80%) of radiolabel in this fraction eluted as a single peak, the putative dolichol-(GlcNAc)2(Man)9(Glc)3 core oligosaccharide. Taken together, these data indicate that the reduction in mannose incorporation by AZT treatment is not due to effects on lipid-linked