Acetylglucosamine-Glycoprotein N-Acetylglucosaminyltransferase

Proc. Nat. Acad. Sci. USA Vol. 72, No. 9, pp. 3323-3327, September 1975 Biochemistry

Chinese hamster ovary cells selected for resistance to the cytotoxicity of phytohemagglutinin are deficient in a UDP-N- acetylglucosamine-glycoprotein N-acetylglucosaminyltransferase activity (lectin-resistant cell mutants/glycosyltransferases/glycoprotein synthesis) PAMELA STANLEY*, SAROJA NARASIMHANt, Louis SIMINOVITCH*, AND HARRY SCHACHTERt * Department of Medical Genetics, University of Toronto, Toronto M5S 1A8 Canada; and t Department of Biochemistry, University of Toronto, Toronto M5S lA8 Canada Communicated by Saul Roseman, June 3, 1975

ABSTRACT Several clones of Chinese hamster ovary teins destined to become binding sites for PHA (11, 12). Evi- cells resistant to the cytotoxicity of the phytohemagglutinin dence for such an hypothesis is presented in this paper; a from Phaseolus vulgaris show decreased binding o 125I-la- preliminary report has appeared (13). beled phytohemagglutinin and contain decreased levels of a UDP-N-acetylglucosamine-glycoprotein N-acetylglucosaminyltransferase (EC 2.4.1.51; UDP-2-acetamido-2-deoxy-D-glu- MATERIALS AND METHODS cose:glycoprotein 2-acetamido-2-deoxy-D-glucosyltransferase) activity when compared to wild-type cells. The decrease in Chemicals and Enzymes. Purified PHA was obtained transferase activity varies from 45% to 96%, depending on from Burroughs Wellcome, England; Con A from Pharma- the exogenous acceptor used in the enzyme assay. No differ- cia Fine Chemicals, Sweden; wheat germ agglutinin (WGA) ences between lectin-resistant and wild-type cells were noted for several other glycosyltransferases. The absence of a par- from Sigma Chemical Co., U.S.A.; Ricinus communts toxin ticular N-acetylglucosaminyltransferase in the lectin-resis- (RIC) was a gift from Dr. C. H. Wei, Oak Ridge National tant cells apparently results in defective glycosylation of lec- Laboratory, Tennessee; and Lens culinaris agglutinin (LCA) tin-binding glycoproteins on the cell surface. A phytohem- was prepared by chromatography from Lens culinaris beans agglutinin-resistant clone which shows decreased binding of (14). Radioactive iodine and GDP-[U-'4C]-L-fucose (89 llsIlabeled phytohemagglutinin but does not exhibit the en- mCi/mmol) were obtained from The Radiochemical Cen- zyme deficiency has also been isolated. tre, Amersham, England, and other radioactive nucleotide Purified plant lectins exert numerous biological effects fol- sugars from New England Nuclear as follows: CMP-N-ace- lowing their specific interaction with the carbohydrate tyl-[4,5,6,7,8,9-'4C]neuraminic acid (150-250 mCi/mmol), moieties of glycoproteins at the cell surface (1, 2). Many lec- UDP-[U-'4C]-D-galactose (258 mCi/mmol), GDP-[U-'4C]- tins are cytotoxic for cultured cells and have been used to se- D-mannose (242 mCi/mmol), and UDP-N-acetyl-[1-"4C]- lect lectin-resistant cell lines (3-8). In our laboratory Chinese D-glucosamine (56.5 mCi/mmol); nonradioactive UDP-ga- hamster ovary (CHO) cells resistant to the cytotoxicity of lactose and UDP-GlcNAc were purchased from Sigma, phytohemagglutinin from Phaseolus vulgaris (PHA) have whereas low specific activity CMP-N-acetylneuraminic acid been isolated and shown to behave as authentic somatic cell (15, 16) and GDP-L-fucose (17) were synthesized in this lab- mutants possessing stable alterations at the cell membrane oratory. Human al-acid glycoprotein was prepared and se- (9). Several independently-selected PHA-resistant (PhaR) quentially degraded with glycosidases as described (18-21). clones survive concentrations of PHA 60- to 100-fold the se- Sialidase (22), partially purified 13-galactosidase (20, 23), and lective dose, bind markedly less '25I-labeled PHA at the cell /3-N-acetylglucosaminidase (fraction DE GM2, ref. 23) were membrane than wild-type cells, and exhibit a 4- to 5-fold in- prepared from Cl. perfringens; the fl-galactosidase prepara- crease in sensitivity compared with wild type to the cytotox- tion contained a small amount of f3-N-acetylglucosaminidase icity of concanavalin A (Con A) (9). PhaR clones with these activity (20) and the f3-N-acetylglucosaminidase preparation properties have recently been found to possess few terminal contained appreciable amounts of 13-galactosidase but was and penultimate galactose residues at the cell surface (10). free of activities capable of hydrolyzing p-nitrophenyl-a- The PhaR phenotype behaves recessively in hybrids formed N-acetyl-D-glucosaminide, p-nitrophenyl-fl-N-acetyl-D-ga- between PhaR and wild-type CHO clones (9). A molecular lactosaminide, and p-nitrophenyl-a-N-acetyl-D-galactosam- basis for the PhaR phenotype consistent with these properties inide. Mr. David Williams of this laboratory purified a gly- might be the absence of an N-acetyl-D-glucosaminyltrans- copeptide from Pronase-digested sialidase-, ,B-galactosidase-, ferase (GlcNAc-T) which attaches d3-linked GlcNAc to a- ,3-N-acetylglucosaminidase-treated al-acid glycoprotein mannosyl termini of growing membrane-bound glycopro- [AGP(-SA,Gal,GlcNAc)] by chromatography on Sephadex G-25, Bio-Gel P-10, and Dowex-5OX2 (pyridinium); this gly- Abbreviations: PHA, phytohemagglutinin from Phaseolus vulgaris; copeptide was shown to have terminal a-mannosyl residues Con A, concanavalin A; WGA, wheat germ agglutinin; RIC, toxin by demonstrating release of mannose on treatment with a- from Ricinus communis; LCA, agglutinin from Lens culinaris; mannosidase from jack bean meal (Boehringer). An immu- GlcNAt-T, N-acetylglucosaminyltransferase; AGP(-SA), sialidase- noglobulin G glycopeptide (24) that had been treated with treated al-acid glycoprotein; AGP(-SA,Gal), sialidase-, f-galactosid- 13-galactosidase and 13-N-acetylglucosaminidase to expose ase-treated al-acid glycoprotein; AGP(-SA,Gal,GlcNAc), sialidase-, f3-galactosidase-, jS-N-acetylglucosaminidase-treated a,-acid glyco- terminal a-mannosyl residues was the gift of Dr. Stuart protein; IgG(-SA,Gal,GlcNAc) glycopeptide, fl-galactosidase-, f3-N- Kornfeld. RNase B from bovine pancreas was purchased acetyl-glucosaminidase-treated glycopeptide from sialic acid-defi- from Worthington. All other chemicals were obtained com- cient immunoglobulin G; CHO, Chinese hamster ovary. mercially. 3323 Downloaded by guest on September 23, 2021 3324 Biochemistry: Stanley et al. Proc. Nat. Ac'ad. Sci. USA 72 (1975) Cell Culture and Cell Lines. The selection and genetic Glycosyltransferase Assays. The incubation mixtures for characterization of PhaR CHO cells has been described in the various transferase assays all contained 0.05-0.14 mg of detail (9). In most of the experiments reported here, com- enzyme protein in addition to the following: (i) Sialyltrans- parisons are made between a parental wild-type CHO clone ferase (EC 2.4.99.1): AGP(-SA), 1 mg; piperazine-N,N'-bis(2- that is auxotrophic for proline (Pro-5) and a previously un- ethanesulfonate) buffer, pH 7.3, 2.5 Mmols; CMP-N-acetyl- described PhaR clone (Pro-PhaR1_1) selected in PHA in a [14C]neuraminic acid, 2.2 X 106 cpm/Mmol, 0.069 gmol; single step from the wild-type line. The clone Pro-5PhaR12- final volume, 0.040 ml. (ii) Fucosyltransferase (EC 2 was selected in PHA by Mr. Steven Rothstein from Pro-5 2.4.1.68): AGP(-SA,Gal), 1 mg; Tris-HCI, pH 8.0, 2.5 11mol; cells. Other cell lines used here are auxotrophic for glycine, MgCl2, 2.0 /imol; GDP-[14C]fucose, 4.4 X 106 cpm/,umol, adenosine, and thymidine and reverted for the proline 0.038 ,mol; final volume, 0.050 nml. (iii) Galactosyltransfer- marker (Gat-Pro+) and have been described (9). This paren- ase (EC 2.4.1.38): GlcNAc, 1.0 Mmol, or AGP(-SA,Gal), 1 tal (wild-type) line is designated Gat-Pro+2, while the inde- mg; 2-(N-morpholino)ethanesulfonate buffer, pH 5.7, 5.0 pendently selected PhaR clones are referred to as Gat- ,umol; MnCl2, 3.0 umol; UDP-[14C]galactose, 3.1 X 106 Pro+2PhaRl and Gat-Pro+lPhaRl, and a clone of the latter cpm/tumol, 0.050 Mmol; final volume, 0.050 ml. (iv) Manno- is Gat-Pro+1PhaRj-j. Cells were routinely grown in suspen- syltransferase: Tris-maleate, pH 7.4, 0.5 ,umol; KCI, 0.1 sion at 370 in complete a medium (25) containing 10% fetal tmnmol; MnCl2, 0.25 Mmol; MgCl2, O.25 urmol; GDP-[I4C]man- calf serum. nose, 6.4 X 107 cpm/,umol, 2.5 nmol; final volume, 0.050 ml. Determination of Lectin Resistance. The abilities of (v) GlcNAc-T (EC 2.4.1.51): AGP(-SA,Gal,GlcNAc), 1 mg, wild-type and PhaR CHO cells to survive cytotoxic levels of or RNase B, 5 mg, or glycopeptide from either AGP- different lectins were compared by determining the dose of (-SA,Gal,GlcNAc) or IgG(-SA,Gal,GlcNAc), 0.1 tumol; 2-(N- lectin (,gg/ml) that reduced the relative plating efficiency morpholino)ethanesulfonate buffer, pH 6.3, 2.5 ,umol; (9) of the cells to 10% (D1o). Lectin resistance was sometimes MnCl2, 0.5 ,umol; UDP-N-acetyl-[14C]glucosamine, 3.5 X determined semiquantitatively by comparing the ability of 106 cpm/,gmol, 0.027 IAmol; final volume, 0.04 ml. Incuba- 2000 cells to form a monolayer at different lectin concentrations were at 370 for 1 hr. The mannosyltransferase incubations in the wells of a Microtest II tray (Falcon Plastics no. tion was terminated by addition of 0.05 ml of ice-cold 10% 3040). In these cases, a lower limit of the Djo value is given. trichloroacetic acid/2% phosphotungstic acid; the precipi- Binding of Iodinated Lectins. PHA, WGA, LCA, and tate was collected on a glass fiber filter and washed three Con A were iodinated with 125I using lactoperoxidase and times with 10 ml of 5% trichloroacetic acid/1% phos- hydrogen peroxide (9). Each labeled lectin preparation was photungstic acid and once with 25 ml of chloroform/metha- 80-90% precipitable by cold 10% (w/v) trichloroacetic acid. nol/water (1:1:0.3). The filter was dried and counted in a Washed cells (5 X 105) were incubated for 1 hr at room tem- Nuclear Chicago liquid scintillation spectrometer. All other perature with '25I-labeled lectin (about 2-4 ,g containing incubations were terminated by addition of 0.010 ml of 2% about 100,000 cpm) in 0.2 ml of phosphate-buffered saline sodium tetraborate/0.25 M EDTA followed by high voltage containing Ca++ and Mg++ and 1% bovine serum albumin. electrophoresis in 1% sodium tetraborate (17-21); the fucos- Unbound 125I-labeled lectin was removed by centrifugation yl- and N-acetylgiucosaminyltransferase incubations were of the reaction mixture through 2.5% or 5% bovine serum al- then subjected to descending paper chromatography with bumin (9). Cell-bound radioactivity was determined in a 80% ethanol. The origins of the papers were dried, and ra- Nuclear Chicago Autogammna Counter. The binding to cells dioactivity was determined by previously described liquid of each '25I-labeled lectin was inhibited about 90% by the scintillation techniques (17-21). presence of a 500- to 1000-fold excess of unlabeled lectin in All incubations were carried out at two protein concentra- the incubation mixture. tions and, except for mannosyltransferase, both in the ab- Preparation of Cell Extracts. Cells in logarithmic growth sence and presence of exogenous acceptors. The incorpora- phase were harvested at a density of 4 to 8 X 105 cells per ml tions in the absence of acceptor were subtracted from exog- and washed three times with cold saline (about 6 ml per 107 enous acceptor values in the calculation of all enzyme veloc- cells). The cell concentration was determined using a Parti- ities. Rate of product formation was proportional to enzyme cle Data Electrozone/Celloscope cell counter, and aliquots protein for all transferases except the mannosyltransferase; containing the required number of washed cells were centri- data for the latter enzyme must therefore be considered pre- fuged at 1000 rpm for 10 min in a MSE Mistral 6 L centri- liminary. fuge. For detergent extraction, 0.025 ml of saline and 0.05 Product Identification. The standard GlcNAc-T incuba- ml of 5% (v/v) Triton X-100 were added per 107 cells to a tions were scaled up 5- to 10-fold and incubated at 370 for 3 final volume of 0.10 ml. The cells were mixed gently at hr. The incubation mixtures were dialyzed against 0.1 M room temperature until lysis had occurred; this was shown NaCl followed by water. The radioactive product was either by phase contrast microscopy to take approximately 30 sec. hydrolyzed with 4 M HCl at 1000 for 4.5 hr or treated with After cooling to 40, undisrupted nuclei and whole cells were 13-N-acetylglucosaminidase in 0.08 M phosphate buffer at removed by centrifugation at 3000 rpm for 1 min as above pH 6.0 for 3 hr and 24 hr at 370 under toluene; these digests and the supernatant was used in the transferase assays. The were analyzed by descending paper chromatography in bu- protein concentrations of wild-type and PhaR Triton X-100 tanol/pyridine/water (45:25:40) and high voltage paper cell extracts were usually between 10 and 14 mg/ml. No electrophoresis in pyridine acetate, pH 3.6. change in the released enzyme activity was observed if the Glycosidase Assays. Incubation mixtures were set up incubation with Triton X-100 was prolonged to 3 hr at 40. identical to the standard GlcNAc-T assays except for the For mannosyltransferase assays, washed cells were suspend- omission of exogenous acceptor and UDP-N-acetyl-[14C]glu- ed in 0.075 ml of saline per 107 cells and disrupted by freez- cosamine and their replacement with 1000-2000 cpm of ra- ing and thawing three times; a supernatant was obtained by dioactive glycoprotein product prepared as above. These re- centrifugation as above and used in the transferase assays. action mixtures were incubated and assayed as-for the stan- The protein concentrations of these extracts were 6-8 mg/ dard transferase assays. ml. Protein. Enzyme protein was determined by the proce- Downloaded by guest on September 23, 2021 Biochemistry: Stanley et a!. Proc. Nat. Acad. Sci. USA 72 (1975) 3325

Table 1. Lectin-resistance and lectin-binding phenotypes of Table 3. Effects of mixing extracts from wild-type (WT) wild-type (Pro-5) and PhaR (Pro-PhaRl-1) CHO cells and PhaR CHO cells on their GlcNAc-T activities % '2I1-Labeled GlcNAc-T activity* D,0* (4g/mi) lectin bound* ProIPhaR (cpm/hr) Pro-PhaR_ Pro-PhaR- 1-1 Exogenous Cell extracts Lectin Pro-5 1-it Pro-5 1-1 Pro-5 acceptor (5-,ul aliquots) Exp. 1 Exp. 2 PHA 3.5 > 1000 13.7 1.1 8 AGP(-SA, WT 3080 2460 Con A 19 3.0 16.8 44.3 264 Gal, GlcNAc) PhaR 1450 1260 WGA 1.5 >60 58.2 26.2 45 WT +PhaR 3770 3760 RIC 0.2 > 3.0 N.D. N.D. LCA 16 > 2000 19.2 9.1 47 Calculated WT + PhaR 4530 3720 * The dose of different lectins that reduced the survival of Pro-5 and Pro-PhaR1-1 cells to 10% and the ability of these cells to bind RNase B WT 930 770 125I-labeled lectins were determined as described in Materials PhaR 330 220 and Methods. N.D. = not determined. WT + PhaR 1880 1130 t The lectin-resistance phenotype of these cells is designated PhaR_ ConASWgaRRicRLcaR. Calculated WT + PhaR 1260 990 dure of Lowry et al. (26), using bovine serum albumin as * Triton X-100 extracts from an equal number of wild-type (Pro-5) standard. Triton X-100 concentrations as high as twice that and PhaR (Pro-PhaRl-1) cells were prepared, and 5-il aliquots present in the enzyme extracts did not interfere with the were assayed alone or in combination for GlcNAc-T activity. protein assays. and Pro-PhaR1-j cell extracts were compared using a vari- ety of exogenous acceptors and nucleotide sugars (Table 2). RESULTS Extracts from PhaR and wild-type cells did not differ signifi- Phenotypic Properties of Wild-Type (Pro-5) and PhaR cantly in their ability to transfer sialic acid, fucose, galac- (Pro-PhaR1_1) CHO Cells. The Pro-PhaR1_1 clone was ob- tose, or mannose to different exogenous or endogenous ac- tained from a colony of Pro- cells which survived plating in ceptors. However, PhaR cell extracts possessed significantly 12 gg/ml of PHA and was maintained thereafter in nonse- less activity than wild-type extracts for the transfer of lective medium. After this single step selection, the clone GlcNAc to terminal a-mannosyl residues. The decrease in was found to be highly resistant not only to PHA, but also to GlcNAc-T activity varied with different exogenous accep- WGA, RIC, and LCA and to be about six times more sensi- tors from 45% with AGP(-SA,Gal,GlcNAc) to 96% with IgG- tive to the cytotoxicity of Con A than wild-type cells (Table (-SA,Gal,GlcNAc) glycopeptide. Endogenous GlcNAc-T ac- 1). Binding studies showed that Pro-PhaR1_1 cells bound tivity was present in both PhaR and wild-type cells and var- only small amounts of 125I-labeled PHA (equivalent to the ied from 1 to 6 nmol/mg of protein per hr with different background of the binding assay, ref. 9), half as much 125I- preparations. The endogenous activity was usually but not labeled WGA and l25-Ilabeled LCA, and about three times always less in PhaR compared with wild-type cell extracts. as much 12I-labeled Con A as wild-type cells (Table 1). The The nuclear pellet routinely removed by centrifugation Pro-PhaR1-1 cells retained the pseudodiploid karyotype of after Triton X-100 extraction (see Materials and Methods) wild-type cells, and their complex phenotype was stable for was found to possess appreciable GlcNAc-T activity, but the more than 200 cell generations. preparation from PhaR cells still possessed only about 50% of Glycosyltransferase Activities of Wild-Type and PhaR the activity in the wild-type pellet. Cells. The glycosyltransferase activities of Pro-5 Kinetic Parameters. The Km values for wild-type and Table 2. Glycosyltransferase activities of wild-type (WT) and PhaR CHO cells Mean glycosyltransferase activities* Glycosyltransferase substrates (nmnol/mg of protein per hr) No. of Nucleotide-sugar Exogenous acceptor experiments WT PhaR % WT

CMP-sialic acid AGP(-SA) 8 5.4 ± 0.90t 6.1 ± 0.91t - GDP-fucose AGP(-SA, Gal) 7 11.6 ± 3.3t 9.5 ± i.lt UDP-galactose GlcNAc 6 18.7 ± 2.6t 19.5 ± 3.1t UDP-galactose AGP(-SA, Gal) 8 9.2 ± 1.3t 9.5 ± 1.5t GDP-mannose None 3 0.28 ± 0.06t 0.27 ± 0.07t UDP-GlcNAc AGP(-SA, Gal, GlcNAc) 15 12.3 ± 3.Ot 6.72 ± 1.8t 55t UDP-GlcNAc RNase B 7 3.7 ± 0.91t 0.94 ± 0.39t 25t UDP-GlcNAc IgG(-SA, Gal, GlcNAc) glycopeptide 2 4.8 0.2 4 UDP-ilcNAc AGP(-SA, Gal, GlcNAc) glycopeptide 2 7.8 3.0 39 * The glycosyltransferase activity of Triton X-100 extracts prepared from washed wild-type (Pro-5) and PhaR (Pro-PhaR1-1) CHO cells were assayed within 30 min of the addition of the detergent by the method described in the text. The arithmetic means 4 standard deviations of the enzyme activities from several separate experiments are presented. t Not significant at P < 0.05. $ Significant at P < 0.001. Downloaded by guest on September 23, 2021 3326 Biochemistry: Stanley et al. Proc. Nat.'Acad. Sci. USA 72 (1975)

Table 4. Substrate specificity of GlcNAc-T activity [14C]GlcNAc or RNase B-["4C]GlcNAc with wild-type or PhaR cell extracts for 1 hr at 370. GlcNAc-T activity Specificity of GlcNAc-T Activity. The GlcNAc-T activity (nmol/mg of from both PhaR and wild-type CHO cells transferred protein per hr) GlcNAc only to receptors with a terminal a-mannosyl resi- Exogenous acceptor WT PhaR due (Table 4). All four exogenous acceptors used ip this study (Table 2) had terminal a-mannosyl residues. AGP(-SA) 0.88 0.72 When either AGP(-SA,Gal,GlcNAc) or RNase B were AGP(-SA, Gal)* 0.60 0.21 used as substrates for PhaR or wild-type cell extracts, acid AGP(-SA, Gal, GlcNAc) 14.7 7.9 hydrolysis of the reaction products released 75-93% of the AGP(-SA, Gal, GlcNAc) radioactivity as free glucosamine, identified by high voltage glycopeptidet 5.25 1.81 paper electrophoresis in pyridine acetate pH 3.6. Treatment x-Mannosidase-treated of these radioactive reaction products with /3-N-acetylgluco- AGP(-SA, Gal, GlcNAc) saminidase at 370 for 24 hr released 80-87% of the radioac- glycopeptidet <0.2 <0.2 tivity as free GlcNAc, identified by descending paper chromatography; radioactive RNase B obtained after incubation WT, wild type. with the PhaR cell extract was not analyzed by glycosidase * Galactose was removed from this preparation by subjecting treatment because of sialidase-treated a1-acid glycoprotein to sequential treatment technical difficulties associated with with periodate, borohydride, and mild acid hydrolysis, as de- the large amount of protein required in the incubation mix- scribed by Spiro (27). ture. t Composition (residues): GlcNAc, 1.8; Man, 2.9; Asn, 1.0; Thr, The ,B-N-acetylglucosaminidase preparation used in these 1.1; Ser, 0.3; Glu, 0.3; Gly, 0.4; Ala, 0.3. experiments contained no activity capable .of removing a- tTreated with jack bean a-mannosidase (Boehringer) to remove linked GlcNAc or a- or ,3-linked GalNAc residues. Thus it 1.8 of the three mannose residues; mannosidase was removed by chromatography on Dowex-50X2 (pyridinium form). appears that the reaction catalyzed by GlcNAc-T in both PhaR and wild-type cell extracts is the transfer of a GlcNAc PhaR cell extracts were, respectively, 0.48 and 0.51 mM for residue in f/-linkage to a terminal a-mannosyl residue. UDP-GlcNAc, 0.21 and 0.29 mM for AGP(-SA,Gal,GlcNAc), Independently Selected PhaR Clones. Several indepen- and 8.0 and 4.8 mM for RNase B. The GlcNAc-T activities dently selected GatrPro+PhaR clones known to bind mark- using either AGP(-SA,Gal,GlcNAc) or RNase B as acceptor edly less 125I-labeled PHA than wild-type cells and to be showed similar pH dependencies (optimum between 6 and sensitive to the cytotoxicity of Con A (9) were also found to 7) and similar Mn++ dependencies (optimum around 12 lack the GlcNAc-T activity (Table 5). However, not all PhaR mM) for both wild-type and PhaR cell extracts. clones exhibited a deficiency in this enzyme activity. One Effect of Mixing Wild-Type and PhaR Extracts on Their clone selected from Pro-5 cells which survived plating in 15 GlcNAc-T Activities. The possibility that PhaR cells may ,ug of PHA per ml was highly resistant to PHA (Djo > 2000 contain GlcNAc-T inhibitor(s) which could account for their ,ug/ml) but possessed GlcNAc-T activity similar to wild-type apparent deficiency in GlcNAc-T activity was investigated (Table 5). This clone (Pro5PhaR12-2) exhibited sensitivities by mixing experiments (Table 3). When PhaR and wild-type similar to wild-type cells for the cytotoxicities of Con A, cell extracts were mixed, the total enzyme activity increased WGA, RIC, and LCA and bound only 20-24% as much to approximately the expected value, indicating that the 125I-labeled PHA as wild-type cells. It is clear that the abili- PhaR cell extract did not inhibit the activity of the wild-type ty to bind 125I-labeled PHA is not directly correlated with a extract. decrease in GlcNAc-T activity in extracts of independent Both PhaR and wild-type cells were assayed directly for PhaR isolates and that a number of complex phenotypes are glycosidase activities by incubation of cell extracts with the associated with PHA resistance in CHO cells. radioactive glycoprotein products of the transferase (see Ma- terials and Methods) under the conditions of transferase DISCUSSION assay. Complete recovery of high-molecular-weight product We have shown that several clones of PhaR CHO cells that was observed after incubating either AGP(-SA,Gal,GlcNAc)- possess the specific lectin-resistance phenotype PhaR_ Table 5. GlcNAc-T activities of independent PhaR and wild-type (WT) CHO clones Phenotypes* GlcNAc-T activities (% WT)t Lectin resistance 125I-labeled PHA bound AGP(-SA, RNase IgG(-SA, GalGlcNAc) Cell lines PHA Con A (% WT) Gal, GlcNAc) B glycopeptide Pro-5 WT WT 100 100 100 100 Pro-PhaR 1-1 R S 10-20 59 17 0 Pro-5PhaR 12-2 R WT 20-24 96 99 N.D. Gat-Pro+2 WT WT 100 100 100 100 Gat-Pro+2PhaRj R S 10 54 28 3 Gat-Pro+lPhaR1 R S 30-40 51 22 0 Gat-Pro+lPhaR 1-1 R S 30-40 37 22 2 * Also presented is a summary of the lectin-resistance and 125I-labeled PHA binding phenotypes of the different clones (see text). R = highly resistant compared with wild type; S = 4- to 6-fold more sensitive than wild type; N.D. = not determined. t Independently selected PhaR clones were extracted with Triton X-100 and assayed for GlcNAc-T activity with various exogenous acceptors. Results are expressed as % of wild-type transferase activities; absolute specific activities were similar to those in Table 2. Downloaded by guest on September 23, 2021 Biochemistry: Stanley et al. Proc. Nat. Acad. Sci. USA 72 (1975) 3327 ConASWgaRRicRLcaR and a decreased ability to bind 'al- cosyltransferase activities of these mutants are now being in- labeled PHA exhibit decreased levels of GlcNAc-T activity vestigated. The PhaR CHO cells and other lectin-resistant towards four exogenous acceptors. The enzyme activity in lines should enable the identity and linkage specificities of both wild-type and PhaR cell extracts appears to catalyze the many mammalian glycosyltransferases to be described. Also transfer of a GlcNAc residue in fl-linkage to a terminal a- they should provide excellent biological material for studies mannosyl residue in the acceptor. The fact that the decrease in cell membrane structure and function. in transferase activity in PhaR clones varies from 40 to 100% We thank Ms. Nancy Stokoe for excellent technical assistance. depending on the exogenous acceptor indicates that PhaR This research was supported by grants from the Medical Research cells are not completely deficient in GlcNAc-T activity but Council of Canada to L.S. and H.S. and from the National Cancer have lost one of two or more such activities present in wild- Institute of Canada and from NIH (U.S.A.) to L.S. P.S. is a postdoc- type cells. Since PhaR cell extracts appear to have essentially toral fellow of the Medical Research Council of Canada. Some of no activity for the transfer of GlcNAc to IgG(-SA,Gal,Glc- the al-acid glycoprotein used in this study was provided by the NAc) glycopeptide, it is likely that PhaR cells lack the activi- American Red Cross National Fractionation Center with the partial ty of a transferase that attaches GlcNAc to mannose via a support of National Institutes of Health Grant HE 13881 (HEM). specific ,B-linkage. Control studies make it unlikely that the decreased enzyme activity in PhaR cells is due primarily to a 1. Lis, H. & Sharon, N. (1973) Annu. Rev. Biochem. 42, 541- shift in the pH optimum or cation requirement of the wild- 573. type transferase or to the presence in PhaR cell extracts of 2. Nicolson, G. L. (1974) Int. Rev. Cytol. 39,89-189. inhibitors (such as a glycosidase or nucleotidase) which 3. Ozanne, B. & Sambrook, J. (1971) in The Biology of Onco- might interfere with the enzyme assays. genic Viruses, ed. Silvestri, L. G. (North Holland Publishing Co., Amsterdam), pp. 248-257. It is at present not possible to determine the site of the 4. Culp, L. A. & Black, P. H. (1972) J. Virol. 9,611-620. mutation in these PhaR cells. The absence of the GlcNAc-T 5. Wright, J. A. (1973) J. Cell Biol. 56,666-675. activity may be a direct or indirect result of the mutational 6. Guerin, C., Zachowski, A., Prigent, B., Paraf, A., Dunia, I., Di- event. Regardless of the mechanism, presumably the en- awara, M.-A. & Benedetti, E. L. (1974) Proc. Nat. Acad. Sci. zyme lesion results in the inability of PhaR cells to attach USA 71, 114-117. GlcNAc residues via a certain :-linkage to a-mannosyl ter- 7. Hyman, R., Lacorbiere, M., Stavarek, S. & Nicolson, G. (1974) mini in the biosynthesis of cellular glycoproteins including J. Nat. Cancer Inst. 52,963-969. those destined for the CHO cell surface. This would in turn 8. Gottlieb, C., Skinner, A. M. & Kornfeld, S. (1974) Proc. Nat. lead to an inability to complete the galactosyl (and sialyl) Acad. Sci. USA 71,1078-1082. termini PHA binding site (11). Previous studies (11, 9. Stanley, P., Caillibot, V. & Siminovitch, L. (1975) Somatic of the Cell Genet. 1,3-26. 12) suggest that the five lectins used here bind to the oligo- 10. Juliano, R. L. & Stanley, P. (1975) Biochim. Biophys. Acta saccharide portion of the same membrane receptor but at 389,401-406. different sites specified partly by their sugar specificities. 11. Kornfeld, R. & Kornfeld, S. (1970) J. Biol. Chem. 245, 2536- Thus, the lack of certain sialyl, l-D-galactosyl and f3-N-ace- 2545. tyl-D-glucosaminyl residues on the PhaR cell surface proba- 12. Toyoshima, S., Fukuda, M. & Osawa, T. (1972) Biochemistry bly accounts for their resistance to PHA, WGA, LCA, and 11,4000-4005. RIC, for their decreased ability to bind l25-Ilabeled PHA, 13. Narasimhan, S., Stanley, P., Siminovitch, L. & Schachter, H. WGA, LCA, and RIC (unpublished observations), and for (1975) Fed. Proc. 34,679. the decreased ability of PhaR cells to be labeled with galac- 14. Howard, I. K., Sage, H. J., Stein, M. D., Young, N. M., Leon, tose borohydride (10); the exposure of man- M. A. & Dyckes, D. F. (1971) J. Biol. Chem. 246,1590-1595. oxidase-tritiated 15. Comb, D. G. & Roseman, S. (1960) J. Biol. Chem. 235,2529- nosyl residues on the PhaR cell surface probably accounts for 2537. the increased binding of l25I-labeled Con A and for the in- 16. Kean, E. L. (1970) J. Biol. Chem. 245,2301-2308. creased sensitivity of PhaR cells to the cytotoxicity of Con A. 17. Jabbal, I. & Schachter, H. (1971) J. Biol. Chem. 246, 5154- However, an additional PhaR phenotype which possesses full 5161. GlcNAc-T activity and a different lectin-resistance pheno- 18. Schachter, H., Jabbal, I., Hudgin, R. L., Pinteric, L., McGuire, type has been described (Table 5). Further, the interpreta- E. J. & Roseman, S. (1970) J. Biol. Chem. 245, 1090-1100. tion of l25-Ilabeled PHA binding appears to be complex; for 19. Hudgin, R. L. & Schachter, H. (1971) Can. J. Biochem. 49, example, clones which bind 30-40% as much l25I-labeled 829-837. PHA as wild-type cells (Gat-Pro+lPhaRl and Gat-Pro- 20. Hudgin, R. L. & Schachter, H. (1971) Can. J. Biochem. 49, do not possess 30-40% as much GlcNAc-T activ- 838-846. +lPhaRlj-) 21. Hudgin, R. L. & Schachter, H. (1971) Can. J. Biochem. 49, ity. Thus it would appear that no single criterion is enough 847-852. to adequately describe the different PhaR phenotypes of 22. Cassidy, J. T., Jourdian, G. W. & Roseman, S. (1965) J. Biol. CHO cells. Chem. 240,3501-3506. The RIC-resistant CHO cells isolated by Gottlieb et al. (8) 23. McGuire, E. J., Chipowsky, S. & Roseman, S. (1972) in Meth- have recently been found to exhibit an almost identical de- ods in Enzymology ed. Ginsburg, V. (Academic Press, New crease in GlcNAc-T activity with IgG(-SA,Gal,GlcNAc) gly- York), Vol. 28, pp. 755-763. copeptide and AGP(-SA,Gal,GlcNAc) (28). We have now se- 24. Kornfeld, R., Keller, J., Baenziger, J. & Kornfeld, S. (1971) J. lected resistant CHO cells using RIC, WGA, and LCA and Biol. Chem. 246,3259-3268. obtained in case clones possessing the first PhaR pheno- 25. Stanners, C. P., Elicieri, G. L. & Green, H. (1971) Nature New each Biol. 230,52-54. type (PhaRConASWgaRRicRLcaR and low levels of GlcNAc- 26. Lowry, 0. H., Rosebrough, N. J., Farr, A. L. & Randall, R. J. T activity) as well as clones with alternative lectin-resistance (1951) J. Biol. Chem. 193,265-275. phenotypes (Stanley et al., manuscript in preparation). It 27. Spiro, R. G. (1964)J. Biol. Chem. 239,567-573. would appear that a large number of phenotypes may be ob- 28. Gottlieb, C., Baenziger, J. & Kornfeld, S. (1975) J. Biol. Chem. tained by these single-step selection techniques, and the gly- 250,3303-309. Downloaded by guest on September 23, 2021