Bovine UDP-N-acetylglucosamine:Lysosomal-enzyme N-Acetylglucosamine-1-phosphotransferase

Abstract

The kinetic properties of UDP-N-acetylglucosamine:lysosomal-enzyme N-acetylglucosamine-1-phosphotransferase (GlcNAc-phosphotransferase) purified to homogeneity from lactating bovine mammary gland have been investigated. GlcNAc-phosphotransferase transferred GlcNAc 1-phosphate from UDP-GlcNAc to the synthetic acceptor α-methylmannoside, generating GlcNAc-1-phospho-6-mannose α-methyl, the structure of which was confirmed by mass spectroscopy. GlcNAc-phosphotransferase was active between pH 5.7 and 9.3, with optimal activity between pH 6.6 and 7.5. Activity was strictly dependent on Mg2+ or Mn2+. The Km for Mn2+ was 185 μM. The Km for UDP-GlcNAc was 30 μM, and that for α-methylmannoside was 63 mM. The enzyme was competitively inhibited by UDP-Glc, with a Ki of 733 μM. The 166-kDa subunit was identified as the catalytic subunit by photoaffinity labeling with azido-[β-32P]UDP-Glc.Purified GlcNAc-phosphotransferase utilizes the lysosomal enzyme uteroferrin ∼163-fold more effectively than the non-lysosomal glycoprotein ribonuclease B. Antibodies to GlcNAc-phosphotransferase blocked the transfer to cathepsin D, but not to α-methylmannoside, suggesting that protein-protein interactions are required for the efficient utilization of glycoprotein acceptors. These results indicate that the purified bovine GlcNAc-phosphotransferase retains the specificity for lysosomal enzymes as acceptors previously observed with crude preparations. The kinetic properties of UDP-N-acetylglucosamine:lysosomal-enzyme N-acetylglucosamine-1-phosphotransferase (GlcNAc-phosphotransferase) purified to homogeneity from lactating bovine mammary gland have been investigated. GlcNAc-phosphotransferase transferred GlcNAc 1-phosphate from UDP-GlcNAc to the synthetic acceptor α-methylmannoside, generating GlcNAc-1-phospho-6-mannose α-methyl, the structure of which was confirmed by mass spectroscopy. GlcNAc-phosphotransferase was active between pH 5.7 and 9.3, with optimal activity between pH 6.6 and 7.5. Activity was strictly dependent on Mg2+ or Mn2+. The Km for Mn2+ was 185 μM. The Km for UDP-GlcNAc was 30 μM, and that for α-methylmannoside was 63 mM. The enzyme was competitively inhibited by UDP-Glc, with a Ki of 733 μM. The 166-kDa subunit was identified as the catalytic subunit by photoaffinity labeling with azido-[β-32P]UDP-Glc. Purified GlcNAc-phosphotransferase utilizes the lysosomal enzyme uteroferrin ∼163-fold more effectively than the non-lysosomal glycoprotein ribonuclease B. Antibodies to GlcNAc-phosphotransferase blocked the transfer to cathepsin D, but not to α-methylmannoside, suggesting that protein-protein interactions are required for the efficient utilization of glycoprotein acceptors. These results indicate that the purified bovine GlcNAc-phosphotransferase retains the specificity for lysosomal enzymes as acceptors previously observed with crude preparations. INTRODUCTIONThe trafficking of lysosomal hydrolases to the lysosome in higher eucaryotes depends on the specific modification of asparagine-linked oligosaccharides to contain a mannose 6-phosphate recognition marker. The initial and determining step in the generation of the mannose 6-phosphate recognition marker is catalyzed by the enzyme UDP-N-acetylglucosamine:lysosomal-enzyme N-acetylglucosamine-1-phosphotransferase (GlcNAc-phosphotransferase). 1The abbreviations used are: GlcNAc-phosphotransferaseUDP-N-acetylglucosamine:lysosomal-enzyme N-acetylglucosamine-1-phosphotransferase5-N3-UDP-Glc5-azidouridine 5′-diphosphoglucoseConAconconavalin AMALDI-TOF-MSmatrix-assisted laser desorption/ionization time-of-flight mass spectroscopyHPLChigh pressure liquid chromatographyDTTdithiothreitolMES4-morpholineethanesulfonic acidBisTris2-[bis(2-hydroxyethyl)amino]]-2-(hydroxymethyl)-propane-1,3-diolPAGEpolyacrylamide gel electrophoresis. In the previous paper (1Bao M. Booth J.L. Elmendorf B.J. Canfield W.M. J. Biol. Chem. 1996; 271: 31437-31445Abstract Full Text Full Text PDF PubMed Scopus (105) Google Scholar), we described the 488,000-fold purification of GlcNAc-phosphotransferase from lactating bovine mammary glands. The purified enzyme was found to be a complex of six subunits and is composed of homodimers of 166-, 56-, and 51-kDa subunits. The identification of GlcNAc-phosphotransferase as a multisubunit enzyme (1Bao M. Booth J.L. Elmendorf B.J. Canfield W.M. J. Biol. Chem. 1996; 271: 31437-31445Abstract Full Text Full Text PDF PubMed Scopus (105) Google Scholar) suggests that it may be possible to link specific properties of the enzyme to specific protein subunits.Previous studies have shown partially purified rat liver GlcNAc-phosphotransferase phosphorylates lysosomal enzymes at least 500-fold better than non-lysosomal glycoproteins with similar high mannose oligosaccharides (2Lang L. Reitman M. Tang J. Roberts R.M. Kornfeld S. J. Biol. Chem. 1984; 259: 14663-14671Abstract Full Text PDF PubMed Google Scholar). Isolated high mannose oligosaccharides were poor substrates, as were heat-denatured lysosomal enzymes (2Lang L. Reitman M. Tang J. Roberts R.M. Kornfeld S. J. Biol. Chem. 1984; 259: 14663-14671Abstract Full Text PDF PubMed Google Scholar). These studies have resulted in a model that proposes that GlcNAc-phosphotransferase recognizes a conformationally sensitive protein determinant found on lysosomal enzymes. The structure of this determinant in the lysosomal enzyme cathepsin D has been characterized (3Baranski T.J. Faust P.L. Kornfeld S. Cell. 1990; 63: 281-291Abstract Full Text PDF PubMed Scopus (105) Google Scholar, 4Baranski T.J. Cantor A.B. Kornfeld S. J. Biol. Chem. 1992; 267: 23342-23348Abstract Full Text PDF PubMed Google Scholar, 5Cantor A.B. Baranski T.J. Kornfeld S. J. Biol. Chem. 1992; 267: 23349-23356Abstract Full Text PDF PubMed Google Scholar). GlcNAc-phosphotransferase partially purified from the soil amoeba Acanthamoeba castellanii is also selective for a protein determinant on lysosomal enzymes (6Ketcham C.M. Kornfeld S. J. Biol. Chem. 1992; 267: 11654-11659Abstract Full Text PDF PubMed Google Scholar).In this paper, we have investigated the enzymatic properties and kinetics of GlcNAc phosphate transfer to lysosomal and non-lysosomal glycoproteins by a homogeneous preparation of bovine GlcNAc-phosphotransferase. Our results demonstrate that bovine GlcNAc-phosphotransferase selectively phosphorylates lysosomal enzymes. That this selectivity, a property previously observed with impure preparations, is also found with isolated GlcNAc-phosphotransferase demonstrates that it is a property of the enzyme itself and not an accessory factor. We also demonstrate, using photoaffinity labeling with 5-N3-[β-32P]UDP-Glc, that the 166-kDa subunit contains the nucleotide sugar-binding site.EXPERIMENTAL PROCEDURESMaterialsScintiverse BD was from Fisher (Pittsburgh, PA). ConA-Sepharose and HiTrap NHS-activated columns were obtained from Pharmacia Biotech Inc. Ribonuclease B and α-methylmannoside were obtained from Sigma. Porcine uteroferrin was the kind gift of Dr. R. Michael Roberts (University of Missouri, Columbia, MO). All other reagents were reagent grade or better and were from standard suppliers.MethodsPreparation of GlcNAc-phosphotransferaseMost of the experiments described were performed with GlcNAc-phosphotransferase that had been purified on PT18-3M-Emphaze according to Method II (1Bao M. Booth J.L. Elmendorf B.J. Canfield W.M. J. Biol. Chem. 1996; 271: 31437-31445Abstract Full Text Full Text PDF PubMed Scopus (105) Google Scholar). In some cases, the purified enzyme was further concentrated by chromatography on Mono Q. All these preparations had a specific activity of 10-12 μmol/mg/h and were homogeneous when examined on silver-stained SDS-polyacrylamide gels. The photoaffinity labeling experiments were performed with GlcNAc-phosphotransferase purified according to Method I (1Bao M. Booth J.L. Elmendorf B.J. Canfield W.M. J. Biol. Chem. 1996; 271: 31437-31445Abstract Full Text Full Text PDF PubMed Scopus (105) Google Scholar). This preparation had a specific activity of 200 nmol/mg/h and was heterogeneous on silver-stained SDS-polyacrylamide gels.Large-scale Preparation of the Enzymatic ProductUDP-GlcNAc (75 nmol) was incubated with 15 μmol of α-methylmannoside in the presence of 50 mM Tris-HCl, pH 7.4, 5 mM MgCl2, and 5 mM MnCl2. GlcNAc phosphotransferase (10 μg, 100,000 units) was added, and the reaction was incubated at 37°C for 18 h. A second reaction containing, in addition to the other components, 1 μCi of [β-32P]UDP-GlcNAc was processed in parallel to calibrate the columns used for product isolation. The reaction was stopped by the addition of 1.5 ml of 5 mM EDTA, pH 8.0, and applied to a 1-ml column of QAE-Sephadex A-25 equilibrated with 2 mM Tris base. The column was washed with 2 volumes of equilibration buffer, and the product was eluted with 2 ml of buffer containing 30 mM NaCl. The product was made 1 M in NH4COOCH3 and desalted on a 1.6 × 50-cm column of Bio-Gel P-2 equilibrated with water. The pooled product was concentrated by rotary evaporation and subjected to mass spectroscopy.Product Identification by Matrix-assisted Laser Desorption/Ionization Time-of-Flight Mass Spectroscopy (MALDI-TOF-MS)The mass of the enzymatic product was determined by MALDI-TOF-MS on a Hewlett-Packard LDI 1700XP mass spectrometer, which was operated at an accelerating voltage of 30 kV, an extractor voltage of 9 kV, and a pressure of 1.7 × 10−6 torr. Polarity was positive. The dried product was dissolved in 100 mM 2,5-dihydroxybenzoic acid, 90% (v/v) methanol. Approximately 1 μl of the sample/matrix was placed on the probe tip and vacuum-crystallized. Samples were desorbed/ionized from the probe tip with a nitrogen laser (λ = 337 nm) with a pulse width of 3 ns and delivering ∼10.5 μJ of energy/laser pulse. The mass spectrum was recorded over a mass to charge (m/z) range of 1,000-2,500. The spectrum was averaged over 27 laser shots and plotted as arbitrary intensity versus m/z.Assay for GlcNAc-phosphotransferaseExperiments using α-methylmannoside as acceptor were performed as described (7Reitman M.L. Lang L. Kornfeld S. Methods Enzymol. 1984; 107: 163-172Crossref PubMed Scopus (24) Google Scholar), except that the [β-32P]UDP-GlcNAc was isolated by HPLC as described (1Bao M. Booth J.L. Elmendorf B.J. Canfield W.M. J. Biol. Chem. 1996; 271: 31437-31445Abstract Full Text Full Text PDF PubMed Scopus (105) Google Scholar), and the radioactivity in each assay was increased to 0.5 μCi. In experiments utilizing glycoprotein acceptors, reaction mixtures containing 0.5-1.0 μCi of [β-32P]UDP-GlcNAc were applied to a 1.0-ml column of ConA-Sepharose previously equilibrated with Tris-HCl, pH 7.6, 1 mM CaCl2, 1 mM MgCl2, and 0.3% Lubrol. Following washing with 20 ml of buffer, the resin was extruded into a scintillation vial and counted with 12 ml of Scintiverse BD. To avoid exceeding the capacity of the ConA-Sepharose columns in experiments using high concentrations of glycoprotein acceptors, only a fraction of the reaction mixture was applied to the ConA-Sepharose column.For experiments using UDP-Glc as donor, [β-32P]UDP-Glc was synthesized similarly to [β-32P]UDP-GlcNAc (1Bao M. Booth J.L. Elmendorf B.J. Canfield W.M. J. Biol. Chem. 1996; 271: 31437-31445Abstract Full Text Full Text PDF PubMed Scopus (105) Google Scholar) and purified by HPLC. The product had a specific activity of 4 mCi/μmol and was stored in 50% ethanol at −20°C.Dependence of GlcNAc-phosphotransferase Activity on Reaction pHReaction mixtures contained 5 mM MgCl2, 5 mM MnCl2, 1 mg/ml bovine serum albumin, 1 mM DTT, and 150 μM [β-32P]UDP-GlcNAc. Each reaction was buffered to the indicated pH with a 250 mM concentration of one of the following buffers: sodium acetate, pH 3.9-4.25; MES/NaOH, pH 4.9-6.4; BisTris-HCl, pH 6.0-6.9; and Tris-HCl, pH 6.9-9.4.Assay of N-Acetylglucosamine-1-phosphodiester α-N-Acetylglucosaminidase (Uncovering Enyzme)[3H]GlcNAc-1-phospho-6-mannose α-methyl was prepared at a specific activity of 2.1 μCi/μmol as described (8Mullis K.G. Ketcham C.M. Anal. Biochem. 1992; 205: 200-207Crossref PubMed Scopus (12) Google Scholar). Uncovering enzyme activity was determined using 1 μg of purified GlcNAc-phosphotransferase as described (8Mullis K.G. Ketcham C.M. Anal. Biochem. 1992; 205: 200-207Crossref PubMed Scopus (12) Google Scholar).Preparation of Glycoprotein SubstratesUteroferrin was used without further purification. Bovine pancreatic ribonuclease B was obtained commercially and chromatographed on ConA-Sepharose to remove contaminating ribonuclease A as described previously (2Lang L. Reitman M. Tang J. Roberts R.M. Kornfeld S. J. Biol. Chem. 1984; 259: 14663-14671Abstract Full Text PDF PubMed Google Scholar). Porcine cathepsin D was purified from frozen spleens as described previously (9Huang J.S. Huang S.S. Tang J. J. Biol. Chem. 1979; 254: 11405-11417Abstract Full Text PDF PubMed Google Scholar). The specific activity was 30 units/mg, and the protein was >90% cathepsin D on silver-stained SDS-polyacrylamide gels.Protein DeterminationProteins were quantitated by absorbance at 280 nm using the following molar extinction coefficients: ribonuclease B, 8,756 (10Sherwood L.M. Potts Jr., J.T. J. Biol. Chem. 1965; 240: 3799-3805Abstract Full Text PDF PubMed Google Scholar); cathepsin D, 20,991 (2Lang L. Reitman M. Tang J. Roberts R.M. Kornfeld S. J. Biol. Chem. 1984; 259: 14663-14671Abstract Full Text PDF PubMed Google Scholar); and uteroferrin, 31,000 (at 545 nm) (11Buhi W.C. Gray W.J. Mansfield E.A. Chun P.W. Ducsay C.A. Bazer F.W. Roberts R.M. Biochim. Biophys. Acta. 1982; 701: 32-38Crossref PubMed Scopus (13) Google Scholar). Protein concentrations for all other proteins including GlcNAc-phosphotransferase were estimated assuming E1 cm1% = 10.0.Photoaffinity Labeling of GlcNAc-phosphotransferase with [β-32P]UDP-Glc5-N3-[β-32P]UDP-Glc was synthesized at a specific activity of 10 mCi/μmol as described previously (12Drake Jr., R.R. Evans R.K. Wolf M.J. Haley B.E. J. Biol. Chem. 1989; 264: 11928-11933Abstract Full Text PDF PubMed Google Scholar, 13Radominska A. Drake R.R. Methods Enzymol. 1994; 230: 330-339Crossref PubMed Scopus (48) Google Scholar), purified by chromatography of the reaction mixture on a column of DEAE-cellulose, and stored in the dark at −20°C in methanol. Samples were prepared as indicated in the legend to Fig. 5. Following a 15-s incubation, the indicated samples were irradiated with a hand-held UV lamp with the glass face removed (5,000 microwatts/cm2; Model UVS-11, Ultra-Violet Products) at a distance of 4 cm for 45 s. Reactions were terminated by the addition of 0.1 volume of 100% trichloroacetic acid and incubation for 30 min at 0°C. The precipitated protein was collected by centrifugation at 15,000 × g for 10 min, and the pellets were dissolved in 25 μl of SDS-PAGE sample buffer by boiling for 5 min at 100°C and analyzed on an SDS-6% polyacrylamide gel (14Laemmli U.K. Nature. 1970; 227: 680-685Crossref PubMed Scopus (205998) Google Scholar). The dye front containing the unreacted 5-N3-[β-32P]UDP-Glc was cut off and discarded. The remaining gel was stained with Coomassie Blue, dried, and analyzed by autoradiography.AntibodiesAffinity-purified polyclonal antibodies were prepared against 488,000-fold purified bovine GlcNAc-phosphotransferase. A New Zealand White rabbit was primed by intradermal injection of 20 μg of GlcNAc-phosphotransferase emulsified in Freund's complete adjuvant. Boosting was at 2-week intervals with 20 μg of GlcNAc-phosphotransferase emulsified in Freund's incomplete adjuvant. Blood was obtained from the central artery of the ear; antiserum was prepared; and immunoglobulins were isolated by precipitation with 50% saturation of ammonium sulfate. The pellet was dissolved in a minimal volume of phosphate-buffered saline and dialyzed against phosphate-buffered saline. GlcNAc-phosphotransferase-specific antibodies were isolated by chromatography of the crude immunoglobulin fraction on a HiTrap NHS-activated column (1 ml) upon which 200 μg of GlcNAc-phosphotransferase had been immobilized according to the manufacturer's instructions. The column was then washed with phosphate-buffered saline followed by water and eluted with 1 mM HCl. The eluted affinity-purified antibody was immediately neutralized with 1 M Tris-HCl, pH 9.0, and concentrated to 1 mg/ml in a Centriprep 30 concentrator, made 1 mM in NaN3, and stored at 4°C.DISCUSSIONIn this study, we used a homogeneous preparation of GlcNAc-phosphotransferase isolated from the lactating bovine mammary gland to investigate the enzymatic properties toward various acceptors. The data presented in this paper demonstrate that the bovine enzyme phosphorylates mammalian lysosomal enzymes better than non-lysosomal enzymes with similar oligosaccharide structures, similar to previous findings with partially purified rat (2Lang L. Reitman M. Tang J. Roberts R.M. Kornfeld S. J. Biol. Chem. 1984; 259: 14663-14671Abstract Full Text PDF PubMed Google Scholar, 16Lang L. Couso R. Kornfeld S. J. Biol. Chem. 1986; 261: 6320-6325Abstract Full Text PDF PubMed Google Scholar) and A. castellanii (6Ketcham C.M. Kornfeld S. J. Biol. Chem. 1992; 267: 11654-11659Abstract Full Text PDF PubMed Google Scholar) enzymes. These data extend the previous findings by demonstrating that this selectivity for lysosomal enzymes is a property of the purified GlcNAc-phosphotransferase complex and not the result of other factors or proteins.The selectivity of bovine GlcNAc-phosphotransferase for lysosomal enzymes results from two factors. First, the Km for uteroferrin is ∼56-fold lower than for ribonuclease B. Second, the Vmax for uteroferrin is ∼3-fold greater than for ribonuclease B. Together, these two factors result in a calculated catalytic efficiency ∼162-fold greater for the lysosomal enzyme acceptor. The use of lysosomal enzymes purified from lysosomes as acceptors is associated with inherent difficulties since the oligosaccharides have frequently been truncated by glycosidases present in the lysosome. The single oligosaccharide of uteroferrin is composed predominantly of Man5GlcNAc2 structures, with lesser amounts of Man6GlcNAc2 and Man4GlcNAc2 (17Saunders P.T.K. Renegar R.H. Raub T.J. Baumbach G.A. Atkinson P.H. Bazer F.W. Roberts R.M. J. Biol. Chem. 1985; 260: 3658-3665Abstract Full Text PDF PubMed Google Scholar). Since α-1,2-linked mannoses are absolutely required for an oligosaccharide to function as an acceptor for GlcNAc-phosphotransferase (18Couso R. Lang L. Roberts R.M. Kornfeld S. J. Biol. Chem. 1986; 261: 6326-6331Abstract Full Text PDF PubMed Google Scholar), only the uteroferrin molecules bearing the Man6GlcNAc2 structures can function as acceptors, while the remaining molecules are competitive inhibitors. Because of these limitations in uteroferrin as an acceptor, the true difference between uteroferrin and ribonuclease B is likely to be at least 4-fold greater than determined. Ribonuclease B contains predominantly Man8GlcNAc2 structures and should provide an optimal oligosaccharide to function as an acceptor. The improved catalytic efficiency of ribonuclease B compared with that of α-methylmannoside likely results from the presence of the preferred α-1,2-linked mannose acceptor. Cathepsin D was also an effective substrate as indicated by a Km of 18 μM; however, the extent of labeling was low, likely as a result of the highly truncated oligosaccharides present on most molecules (19Takahashi T. Schmidt P.G. Tang J. J. Biol. Chem. 1983; 258: 2819-2830Abstract Full Text PDF PubMed Google Scholar). The amount of cathepsin D available prevented fractionation into specific glycoforms that would be expected to function as more efficient acceptors.With the identification of bovine GlcNAc-phosphotransferase as a multiple subunit enzyme, the association of specific functions with specific protein subunit(s) becomes possible. UDP-Glc was found to be a competitive inhibitor/alternate substrate for bovine GlcNAc-phosphotransferase, similar to results previously reported for the rat enzyme (7Reitman M.L. Lang L. Kornfeld S. Methods Enzymol. 1984; 107: 163-172Crossref PubMed Scopus (24) Google Scholar). The finding of strictly competitive inhibition indicates that UDP-Glc and UDP-GlcNAc compete for the same nucleotide sugar-binding site. This observation allowed the identification of the subunit containing the nucleotide sugar-binding site with the photoprobe 5-N3-UDP-Glc. The 166-kDa subunit is specifically photoaffinity-labeled with the photoprobe 5-N3-[β-32P]UDP-Glc, indicating that the nucleotide sugar-binding site is localized to this subunit. INTRODUCTIONThe trafficking of lysosomal hydrolases to the lysosome in higher eucaryotes depends on the specific modification of asparagine-linked oligosaccharides to contain a mannose 6-phosphate recognition marker. The initial and determining step in the generation of the mannose 6-phosphate recognition marker is catalyzed by the enzyme UDP-N-acetylglucosamine:lysosomal-enzyme N-acetylglucosamine-1-phosphotransferase (GlcNAc-phosphotransferase). 1The abbreviations used are: GlcNAc-phosphotransferaseUDP-N-acetylglucosamine:lysosomal-enzyme N-acetylglucosamine-1-phosphotransferase5-N3-UDP-Glc5-azidouridine 5′-diphosphoglucoseConAconconavalin AMALDI-TOF-MSmatrix-assisted laser desorption/ionization time-of-flight mass spectroscopyHPLChigh pressure liquid chromatographyDTTdithiothreitolMES4-morpholineethanesulfonic acidBisTris2-[bis(2-hydroxyethyl)amino]]-2-(hydroxymethyl)-propane-1,3-diolPAGEpolyacrylamide gel electrophoresis. In the previous paper (1Bao M. Booth J.L. Elmendorf B.J. Canfield W.M. J. Biol. Chem. 1996; 271: 31437-31445Abstract Full Text Full Text PDF PubMed Scopus (105) Google Scholar), we described the 488,000-fold purification of GlcNAc-phosphotransferase from lactating bovine mammary glands. The purified enzyme was found to be a complex of six subunits and is composed of homodimers of 166-, 56-, and 51-kDa subunits. The identification of GlcNAc-phosphotransferase as a multisubunit enzyme (1Bao M. Booth J.L. Elmendorf B.J. Canfield W.M. J. Biol. Chem. 1996; 271: 31437-31445Abstract Full Text Full Text PDF PubMed Scopus (105) Google Scholar) suggests that it may be possible to link specific properties of the enzyme to specific protein subunits.Previous studies have shown partially purified rat liver GlcNAc-phosphotransferase phosphorylates lysosomal enzymes at least 500-fold better than non-lysosomal glycoproteins with similar high mannose oligosaccharides (2Lang L. Reitman M. Tang J. Roberts R.M. Kornfeld S. J. Biol. Chem. 1984; 259: 14663-14671Abstract Full Text PDF PubMed Google Scholar). Isolated high mannose oligosaccharides were poor substrates, as were heat-denatured lysosomal enzymes (2Lang L. Reitman M. Tang J. Roberts R.M. Kornfeld S. J. Biol. Chem. 1984; 259: 14663-14671Abstract Full Text PDF PubMed Google Scholar). These studies have resulted in a model that proposes that GlcNAc-phosphotransferase recognizes a conformationally sensitive protein determinant found on lysosomal enzymes. The structure of this determinant in the lysosomal enzyme cathepsin D has been characterized (3Baranski T.J. Faust P.L. Kornfeld S. Cell. 1990; 63: 281-291Abstract Full Text PDF PubMed Scopus (105) Google Scholar, 4Baranski T.J. Cantor A.B. Kornfeld S. J. Biol. Chem. 1992; 267: 23342-23348Abstract Full Text PDF PubMed Google Scholar, 5Cantor A.B. Baranski T.J. Kornfeld S. J. Biol. Chem. 1992; 267: 23349-23356Abstract Full Text PDF PubMed Google Scholar). GlcNAc-phosphotransferase partially purified from the soil amoeba Acanthamoeba castellanii is also selective for a protein determinant on lysosomal enzymes (6Ketcham C.M. Kornfeld S. J. Biol. Chem. 1992; 267: 11654-11659Abstract Full Text PDF PubMed Google Scholar).In this paper, we have investigated the enzymatic properties and kinetics of GlcNAc phosphate transfer to lysosomal and non-lysosomal glycoproteins by a homogeneous preparation of bovine GlcNAc-phosphotransferase. Our results demonstrate that bovine GlcNAc-phosphotransferase selectively phosphorylates lysosomal enzymes. That this selectivity, a property previously observed with impure preparations, is also found with isolated GlcNAc-phosphotransferase demonstrates that it is a property of the enzyme itself and not an accessory factor. We also demonstrate, using photoaffinity labeling with 5-N3-[β-32P]UDP-Glc, that the 166-kDa subunit contains the nucleotide sugar-binding site.