and neutral type [(mannose)5_- N- charge heterogeneity, with pl values ranging from 4.2 to 5.0. However, the possible presence

Transcription

1 JOURNAL OF VIROLOGY, Jan. 1983, p Vol. 45, No X/83/ $02.00/0 Copyright 1982, American Society for Microbiology Comparison of the Oligosaccharide Moieties of the Major Envelope Glycoproteins of the Subgroup A and Subgroup B Avian Myeloblastosis-Associated Viruses LAWRENCE A. HUNT`* AND STEPHEN E. WRIGHT2'3 Department of Microbiology and Immunology, University of Louisville School of Medicine, Louisville, Kentucky ; Viral Oncology Laboratory, Veterans Administration Medical Center, Salt Lake City, Utah ; and Departments of Medicine and Cellular, Viral and Molecular Biology, University of Utah College of Medicine, Salt Lake City, Utah Received 21 July 1982/Accepted 30 September 1982 The nature of the oligosaccharide chains of the major envelope glycoprotein, gp85, from avian myeloblastosis-associated viruses has been examined for the subgroup A and subgroup B viruses replicated in fibroblasts from the same chicken embryos. Pronase-digested glycopeptides from [3H]mannose- or [3H]glucosamine-labeled viruses were analyzed by the combined techniques of gel filtration, endo-p-n-acetylglucosaminidase digestion, and concanavalin A affinity chromatography. The gp85 protein from these two viruses, and also from another subgroup A avian leukosis virus replicated in the same cells, contained a diverse array of asparagine-linked oligosaccharides of the acidic type [(sialic acid±galactose-n - acetylglucosamine)2-4- (mannose)3 - N - acetylglucosamine2( ±fucose)- asparagine], hybrid type (sialic acid±galactose-n-acetylglucosamine-(mannose)5,4-n-acetylglucosaminez-asparagine), and neutral type [(mannose)5_- N- acetylglucosamine2-asparagine], with the more highly branched (tri- or tetraantennary or both) acidic-type structures representing the predominant class of oligosaccharide. Minor differences were observed between the gp85 of the subgroup B versus subgroup A viruses. The standard strain of avian myeloblastosis virus (AMV), BAI strain A, contains both a replication-defective virus with in vitro hematopoietic cell transforming activity and nondefective helper viruses of subgroups A and B that are designated myeloblastosis-associated viruses 1 and 2 (MAV-1, subgroup A; MAV-2, subgroup B) (14, 21). MAV-1 and MAV-2 have been shown to induce osteopetrosis, nephroblastoma, and lymphoid leukosis in infected chickens, even though neither virus has in vitro transforming abilities (1, 21, 25). Whereas both of these helper viruses can replicate in chicken embryo fibroblasts (CEF), only the subgroup B virus is able to replicate in macrophages and myeloblasts (21). Only a fraction of the envelope glycoprotein apparently determines the host range and other subgroup-specific properties, since the oligonucleotide maps of genome RNA from MAV-1 and MAV-2 are very similar except for a difference in three of the oligonucleotides from the envelope region (3, 24). A number of characteristics of the major glycoprotein, gp85, of the standard strain of AMV have been determined (19, 20, 23). (i) the apparent molecular weight from polyacrylamide gel electrophoresis is approximately 80,000; (ii) the 233 carbohydrate content is approximately 40% by weight and consists of N-acetylglucosamine (GlcNAc), galactose (Gal), mannose (Man), sialic acid (NeuNAc), and fucose (in decreasing order of abundance), but not N-acetylgalactosamine; and (iii) the protein exhibits significant charge heterogeneity, with pl values ranging from 4.2 to 5.0. However, the possible presence of both subgroup A and B virus glycoproteins in unknown ratios and the undetermined cellular sources(s) of virus isolated from the sera of infected chickens make the interpretation of these results difficult. The host cell may be especially important in the analysis of the carbohydrate moieties of AMV or MAV-2 glycoproteins: the oligosaccharides were antigenically reactive when derived from virus isolated from sera of infected chickens or AMV-infected myeloblasts in culture, but not when derived from virus grown in CEF (27, 28). In addition, the host cell-dependent glycosylation of avian retrovirus glycoproteins is altered in transformed versus nontransformed CEF (10, 17). The objectives of the present studies were to isolate gp85 from 3H-sugar-labeled MAV-1 and MAV-2 replicated in fibroblasts from the same chicken embryos and to analyze and compare

2 234 HUNT AND WRIGHT the radiolabeled oligosaccharide components using the combined techniques of gel filtration, endo-p-n-acetylglucosaminidase digestion, and lectin-affinity chromatography. In addition, these gp85 carbohydrate moieties were compared with those derived from a Rous-associated virus of subgroup A (RAV-1) that was also replicated in nontransformed CEF. MATERIALS AND METHODS Preparation of radiolabeled virus. CEF of C/E phenotype were prepared from embryonated eggs (Life Sciences Inc., St. Petersburg, Fla.) and grown in tissue culture as described previously (10, 13). Cultures derived from the same embryos were infected at the third passage with either MAV-1, MAV-2, or RAV-1. The RAV-1 was obtained from Paul Neiman (University of Washington) (29), and the MAV-1 and MAV-2 were obtained from J. H. Chen (Life Sciences, Inc.). These strains of MAV-1 and MAV-2 were originally isolated in the laboratory of Carlo Moscovici (25). Virus-infected cells at the eighth passage were grown to confluency in 75-cm2 culture flasks, and radiolabeling was done in Eagle minimum essential medium with one-third the normal concentration of glucose (0.33 mg/ml) and supplemented with 2% fetal bovine serum (K. C. Biologicals) and either 200,uCi of [2-3H]mannose (14.5 Ci/mmol; New England Nuclear Corp.) per ml or 50,uCi of [6-3H]glucosamine (38 Ci/ mmol; Amersham Corp.) per ml. After labeling for approximately 24 h at 37 C, the medium was harvested, and fresh medium lacking radioactive sugars, but containing the normal amount of glucose and 2% fetal bovine serum, was added. The cultures were incubated for an additional 24 h at 37 C, and this second medium was also harvested. The two harvests of medium for each virus-infected culture were combined, and radiolabeled virus was purified from clarified medium by ultracentrifugation (13). Sodium dodecyl sulfate-polyacrylamide gel electrophoresis. Analytical gel electrophoresis of samples of purified virus was performed in 10% polyacrylamide slab gels by using the discontinuous buffer system of Laemmli (16). The major viral glycoprotein species, gp85, was isolated by preparative electrophoresis of virus in 8% polyacrylamide cylindrical gels (7). Preparation and gel filtration of pronase-digested glycopeptides. Radiolabeled gp85 was extensively digested with pronase (Calbiochem) (5), and the glycopeptides were desalted on a column of Sephadex G15/ G50 (Pharmacia Fine Chemicals) (11). Glycopeptides and oligosaccharides were analyzed by gel filtration through columns (120 by 1.5 cm) of Bio-Gel P-4 (minus 400 mesh; Bio-Rad Laboratories) along with unlabeled and "C-labeled gel filtration markers (10, 11, 13). The elution positions of a series of neutral oligomannosyl cores of known composition [(Man)"GlcNAcj; n = 3 to 9] were used to calibrate the Bio-Gel P-4 columns before the gel filtration of endo-o-n-acetylglucosaminidase digestion products (8, 10). Glycosidase digestions of glycopeptides. Purified endo-p-n-acetylglucosaminidase D (endoglycosidase D) from Diplococcus pneumoniae and endo-3-n-acetylglucosaminidase H (endoglycosidase H) from Streptomyces griseus were purchased from Miles Laboratories, Inc. Glycopeptides were digested with endoglycosidase D and then endoglycosidase H as previously described (5). Lectin affinity chromatography. Endoglycosidase D- and H-digested glycopeptides were fractionated on columns of concanavalin A (ConA)-agarose (Sigma Chemical Co.) as recently described in detail elsewhere (9). Unbound glycopeptides or oligosaccharides (or both) were washed through the columns with 10 mm Tris (ph 7.4) and bound glycopeptides or oligosaccharides were sequentially eluted with the same buffer containing 10, 50, and 100 mm a-methyl mannoside. Peak fractions of unbound or bound and eluted radiolabel were concentrated by lyophilization. RESULTS Electrophoretic comparisons of the glycoproteins of MAV-1, MAV-2, and RAV-1. The only major [3H]mannose-labeled protein observed in the sodium dodecyl sulfate-polyacrylamide electrophoretic profiles of virus purified from the growth medium of MAV-1-, MAV-2-, or RAV-1- infected CEF cultures was gp85, the major envelope glycoprotein of avian retroviruses (Fig. 1). The electrophoretic mobility of MAV-2 gp85 (Fig. lb) was slightly lower than the mobility of gp85 from the two subgroup A viruses, RAV-1 and MAV-1 (Fig. la and c). In addition, MAV-2 gp85 exhibited greater apparent molecular weight heterogeneity than MAV-1 gp85. Differences in the electrophoretic mobility have been reported earlier between gp85 from avian retro- Aki- MWAg'-AA& a b c J. VIROL. ~4 -gp85 FIG. 1. Sodium dodecyl sulfate-polyacrylamide gel electrophoresis of [3H]mannose-labeled MAV-1, MAV-2, and RAV-1. Purified virus was subjected to electrophoresis on a polyacrylamide slab gel, and bands of radiolabeled protein were detected by fluorography (2). Lanes: (a) RAV-1, (b) MAV-2, (c) MAV- 1. gp85 refers to the major envelope glycoprotein.

3 VOL. 45, cm 3 '0 x 2 2- O I ENVELOPE GLYCOPROTEINS OF AVIAN MAVs 235 viruses of different subgroups (4, 6), with sub- A group B gp85 exhibiting a lower mobility than subgroup A gp85 when the viruses were both released from Rous sarcoma virus (RSV)-transformed CEF (6). Several minor proteins of higher molecular -Il ) weight were detected in both the MAV-1 and MAV-2 samples (Fig. lb and c), and the minor envelope glycoprotein species, gp37, was detected in all three samples after longer fluorographic exposures (data not shown). Because of the presence of these minor glycoprotein species along with the major gp85 species, [3H]man- OAB*2O. nose-labeled gp85 was isolated by preparative B polyacrylamide gel electrophoresis of the three purified viruses before analysis of the oligosaccharide moieties of the viral glycoproteins. Gel filtration of [3H]mannose-labeled glycopeptides and oligosaccharides from MAV-1, MAV-2, -2 8 and RAV-1 gp85. The pronase-digested glycopeptides from gp85 of all three viruses were heterogenous in size (Fig. 2), with the larger glycopeptides exhibiting resistance to the endoglycosidases (Table 1) that is characteristic of acidic-type asparaginyl-oligosaccharides [NeuNAc + Gal-GlcNAc)J(Man)3GlcNAc2( ± fuanayn cose)-asn] (10). The digestion products of en- C doglycosidases D and H (Table 1) included both (Man),9GlcNAcl-sizeoligosaccharides(fractions 70 through 90) and a major peak (fractions 60 through 70) in the position expected for endoglycosidase H released hybrid-type oligosaccharides [NeuNAc-Gal-GlcNAc-(Man)5GlcNAc1] l ) previously demonstrated for Prague C RSV (PrC RSV) gp85 (12). The size distribution of the endoglycosidaseresistant glycopeptides was similar for all three Il viruses and was characteristic of the profiles of [3H]mannose-labeled, acidic-type glycopeptides of a transformation-defective (td) derivative of PrC RSV released from nontransformed CEF (10). In addition, the distribution of radiolabel FRACTION NUMBER among the endoglycosidase-resistant glycopeptides (presumably acidic type), hybrid-type FIG. 2. Bio-Gel P-4 gel filtration of untrea endoglycosidase-treated glycopeptides from ptednfid oligosaccharides, and (Man)5_9GlcNAc1-size gp85 of [3H]mannose-labeled MAV-1, MAVpu2r neutral-type oligosaccharides for MAV-1 and RAV-1. Untreated (0) and endoglycosidase D- and H- MAV-2 gp85 was almost identical to the distri- bution reported earlier for td PrC RSV glycopep- digested glycopeptides (0) were separately alnalyzed on the same column, and the profiles of racdiolabel tides: 53% acidic, 9% hybrid, and 38% neutral were superimposed by alignment of the peak elution for MAV-1 (Fig. 2A); 56% acidic, 10% hybrid, positions of the gel filtration standards. The thriee solid and 34% neutral for MAV-2 (Fig. 2B); and 57% vertical arrows represent, from left to right, tihe peak acidic, 8% hybrid, and 35% neutral for td PrC elution positions of blue dextran (void volume) 1, stachv repre- yose, and mannose; the dashed vertical arro% sents the peak elution position of a [14C]glucorsamine- and 5) refer to the elution positions of neutral oligolined numbers (7, 6, 5) in panel A (and also in Fig. 4 labeled (Man)5GlcNAcj neutral oligosacccharide derived from the endoglycosidase D digestion of gly- Chi- respectively) in the endoglycosidase digestion prodmannosyl cores [(Man)nGlcNAcj, with n = 7, 6, 5, copeptides from a phytohemmagglutinin-resistaant nese hamster ovary cell line (8). (A) Glycop)eptides ucts, and x refers to the peak of sialic acid-containing from MAV-1 gp85; (B) glycopeptides from MAV-2 hybrid-type oligosaccharides released by endoglycosidase H gp85; (C) glycopeptides from RAV-1 gp85. The under- (12).

4 236 HUNT AND WRIGHT J. VIROL. TABLE 1. Properties of asparaginyl oligosaccharidesa Endoglyco- ConA- Oligosaccharide type sidase H agarose sensitivity affinity Acidic Diantennary, (NeuNAc±Gal-GlcNAc)2(Man)3GIcNAc2(±fucose)-AsN Resistantb Bound Tri- or tetraantennary, (NeuNAc±Gal-GlcNAc)34(Man)3GlcNAC2(+fucose)-AsN Resistantb Unbound Neutral (Man), GlcNAc2-AsN Sensitivec Boundd Hybrid Five mannose (major species), NeuNAc±Gal-GlcNAc-(Man)5GlcNAc2-AsN Sensitive Bound Four mannose (minor species), NeuNAc±Gal-GIcNAc-(Man)4GlcNAc2-AsN Sensitive Unbound a b Data taken from references 9, 10, and 13. These glycopeptides are sensitive to endoglycosidase D only after branch sugars (NeuNAc-Gal-GIcNAc) have been removed by exoglycosidases. c The five-mannose structure [(Man)5GlcNAc2-AsN] is also sensitive to endoglycosidase D. A smaller neutraltype glycopeptide [(Man)3GlcNAc2-AsN] present in viral glycopeptides from transformed cells is resistant to endoglycosidase H, but sensitive to endoglycosidase D (8, 10). d The larger oligomannosyl core structures (containing seven to nine mannoses) are more tightly bound to ConA-agarose and are only eluted at higher concentrations of a-methyl mannoside than required to elute smaller oligomannosyl core structures (containing three, five, or six mannoses) (9). RSV (10). The fraction of radiolabel in endoglycosidase-resistant glycopeptides was slightly increased for the RAV-1 gp85 sample (Fig. 2C): 64% acidic, 8% hybrid, and 28% neutral. Krantz et al. (15) have previously reported that gp85 from RAV-1 replicated in RSV-transformed CEF has both acidic- and neutral-type asparaginyl oligosaccharides in a ratio of approximately 3:1. Minor amounts of radiolabeled gp37 were obtained from the preparative polyacrylamide gel electrophoresis of [3H]mannose-labeled MAV-2 and RAV-1, and the pronase-digested glycopeptides from these proteins were analyzed in the same manner (data not shown). The minor envelope glycoprotein species from both viruses was apparently enriched for acidic-type oligosaccharides, with relatively less hybridand neutral-type oligosaccharides compared with gp85 from the same viruses. Analysis of the oligosaccharide components of gp37 was difficult because: (i) radiolabel in gp37 was recovered in very low amounts compared with gp85 (Fig. 1); and (ii) the lower molecular weight range on preparative polyacrylamide gels might be contaminated with nonspecific degradation fragments of gp85. Fractionation of viral glycopeptides and oligosaccharides by ConA affinity chromatography. The endoglycosidase-digested glycopeptides from [3H]mannose-labeled MAV-1, MAV-2, and RAV-1 gp85 were subjected to fractionation by ConA-agarose affinity chromatography to analyze the microheterogeneity in the acidic-type glycopeptides: diantennary structures [(NeuN- Ac + Gal - GlcNAc)2-(Man)3GlcNAc2(+fucose)- AsN] should be bound to the column and eluted with a-methyl mannoside, whereas the more highly branched tri- or tetraantennary structures [(NeuNAc + Gal - GlcNAc)3-,(Man)3( ± fucose) - AsN] should not be bound (Table 1) (9). In addition, the endoglycosidase-released oligosaccharides could be further characterized by their affinity for ConA (Table 1): (i) hybrid-type oligosaccharides with a five-mannose core should bind, whereas hybrid-type oligosaccharides with a four-mannose core should not bind (9, 12); and (ii) large neutral oligomannosyl core structures [(Man)7_9GlcNAc1] should be more tightly bound to ConA and less easily eluted with a-methyl mannoside than smaller neutral oligosaccharides [(Man)56GlcNAcj] (9). The profile of radiolabel that was unbound to a column of ConA-agarose versus bound and eluted with a-methyl mannoside is shown in Fig. 3 for the endoglycosidase D- and H-digested glycopeptides from [3H]mannose-labeled MAV- 2 gp85. Approximately 90% of the radiolabel applied to the column was recovered in fractions 1 through 32, and similar profiles and recoveries of radiolabel were obtained with the MAV-1 and RAV-1 gp85 samples (data not shown). The gel filtration patterns of radiolabel in the ConA-unbound versus bound and 10 mm a- methyl mannoside-eluted fractions are displayed in Fig. 4 for the endoglycosidase-digested glycopeptides of all three viruses. The larger average size of the ConA-unbound and endoglycosidaseresistant glycopeptides (Fig. 4) compared with the ConA-bound, eluted glycopeptides (fractions 45 through 60, Fig. 4) was consistent with the former containing more highly branched acidic-type oligosaccharides (tri- or tetraantennary or both) and the latter containing diantennary acidic-type oligosaccharides. The MAV-1 and RAV-1 gp85 glycopeptides contained rela-

5 VOL. 45, x 10 I-) a:ii loomm t I i ENVELOPE GLYCOPROTEINS OF AVIAN MAVs 237 position of (Man)89GlcNAc1-size oligosaccharides (fractions 72 through 80). Because oligosaccharides with larger oligomannosyl cores [(Man)7_9GlcNAc1j have been previously shown to bind tighter to ConA-agarose than smaller oligomannosyl core structures (9) and because asialo forms of the hybrid-type oligosaccharides FRACTION NUMBER FIG. 3. ConA affinity chromatography of [3H]mannose-labeled glycopeptides from MAV-2 gp85 after digestion with endoglycosidases D and H. A portion of endoglycosidase-digested glycopeptides equivalent to those displayed in Fig. 2B was fractionated on a column of ConA-agarose, and 0.1-ml samples of each 1.0-ml fraction were assayed for radioactivity. Bound radiolabel was eluted with increasing concentrations of a-methyl mannoside as indicated by the dotted vertical arrows (10 mm for fractions 9 through 16; 50 mm for fractions 17 through 24; 100 mm for fractions 25 through 32). Peak tubes of unbound material or material that was bound and subsequently eluted with a-methyl mannoside were pooled for further gel filtration analysis (Fig. 4B), as indicated by the bracketed arrows. CY 'O x Q 0L) tively more of the lower-molecular-weight, ConA-unbound glycopeptides (fractions 50 through 60, Fig. 4A and C) than the MAV-2 gp85 glycopeptides (Fig. 4B). The apparent ratio of tri- or tetraantennary (or both) acidic-type glycopeptides to diantennary acidic-type glycopeptides was also different between the glycopeptides of MAV-2 gp85 and the glycopeptides of MAV-1 and RAV-1 gp85: the ratio of radiolabel in ConA-unbound versus ConA-bound, eluted glycopeptides (fractions 37 through 62, Fig. 4) was 1.8:1.0 for MAV-2 gp85, 3.4:1.0 for MAV-1 gp85, and 3.1:10 for RAV-1 gp85. Much of the size heterogeneity in these endoglycosidase-resistant glycopeptides was apparently due to sialic acid heterogeneity, since removal of terminal sialic acid by hydrolysis with 0.1 N H2SO4 at 80 C for 30 min (5) shifted most of the glycopeptides to later eluting fractions and reduced the size heterogeneity (data not shown). The endoglycosidase-released oligosaccharides that were recovered in the 10 mm a-methyl mannoside-eluted fractions from the ConA-agarose columns (fractions 62 through 90, Fig. 4) included a major peak in the position of hybrid-type oligosaccharides with five mannose cores (fractions 63 through 68), a peak of (Man)s 6GlcNAcl-size oligosaccharides (fractions 85 through 90), and another peak in the FRACTION NUMBER FIG. 4. Bio-Gel P-4 gel filtration of [3H]mannoselabeled glycopeptides and oligosaccharides from virion gp85 after fractionation by ConA-agarose chromatography. The conditions of gel filtration were identical to those described in Fig. 2. The endoglycosidase D- and H-digested glycopeptides that were unbound to the lectin affinity column (0; corresponding to fractions 1 through 4 in Fig. 3 for the MAV-2 sample) and the glycopeptides or oligosaccharides that were eluted with 10 mm a-methyl mannoside (0; corresponding to fractions 10 through 16 in Fig. 3) were separately analyzed on the same column, and the proffles were superimposed. (A) MAV-1 gp85 glycopeptides; (B) MAV-2 gp85 glycopeptides; (C) RAV-1 gp85 glycopeptides.

6 238 HUNT AND WRIGHT from PrC RSV glycoproteins elute in the position of (Man)8gGlcNAc1-size oligosaccharides (12), the radiolabel in these (Man)89GlcNAc1- size oligosaccharides most likely represented asialo-hybrid oligosaccharides [Gal-GlcNAc- (Man)5GlcNAcl]. In addition, minor peaks were observed in the ConA-unbound fraction (fractions 66 through 69 and 76 through 79, Fig. 4) eluting just after the peaks of the hybrid-type and presumed asialo-hybrid oligosaccharides from the ConA-bound and eluted fraction. These ConA-unbound structures were consistent with the minor species of hybrid-type oligosaccharides containing a four-mannose core (Table 1) that were previously described for PrC RSV glycoproteins (9, 12). The radiolabel that was eluted from the ConA-agarose columns with 50 and 100 mm a-methyl mannoside (fractions 18 through 31, Fig. 3) included (Man)7_9GlcNAc1- size oligosaccharides and lesser amounts of (Man)_6GlcNAcl-size oligosaccharides for all three virus gp85 samples (data not shown). Analysis of [31H]glucosamine-labeled glycopeptides from MAV-1 and MAV-2 gp85. In further studies of the structural heterogeneity of the acidic-type oligosaccharides of MAV-1 and MAV-2 gp85, [3H]glucosamine-labeled glycopeptides derived from polyacrylamide gel-purified protein were analyzed by the same methods (Table 1). As shown in Fig. 5A for the MAV-2 gp85 sample, the [3H]glucosamine-labeled glycopeptides were greatly enriched for the largermolecular-weight and endoglycosidase-resistant glycopeptides compared with the analogous [3H]mannose-labeled glycopeptides in Fig. 2B. Only minor amounts of radiolabel from the endoglycosidase H-digested glycopeptides eluted in the positions of the sialic acid-containing hybrid-type oligosaccharides (fractions 64 through 68, Fig. 5) and the (Man)5_9GlcNAc1- size oligosaccharides (fractions 70 through 90, Fig. 5). In addition, a minor amount of radiolabel eluted in fractions 90 through 100 between the 14C-labeled Man5GlcNAc1 marker and the stachyose marker, compatible in size with the radiolabeled N-acetylglucosamine peptide product of endoglycosidase H digestion (13). The gel filtration profiles of ConA-unbound versus ConA-bound glycopeptides from [3H]glucosamine-labeled MAV-2 gp85 (Fig. 5B) and MAV-1 gp85 (Fig. SC) were similar to those with the corresponding [3H]mannose-labeled acidictype glycopeptides (Fig. 4A and B) with respect to both size distribution and ratio of radiolabel in ConA-unbound (presumably tri- or tetraantennary or both) versus ConA-bound, eluted (presumably diantennary) acidic-type structures (2.4:1.0 for MAV-2; 3.3:1.0 for MAV-1). A minor peak of hybrid-size oligosaccharides was recovered in the ConA-bound and 10 mm a- 06 (MAV- 2) 46 (MAVy-I) 2 J. VIROL FRACTION NUMBER FIG. 5. Bio-Gel P-4 gel filtration of glycopeptides and oligosaccharides from purified gp85 of [3H]glucosamine-labeled MAV-1 and MAV-2. The conditions of gel ifitration were identical to those in Fig. 2. (A) Superimposed profiles of untreated (0) and endoglycosidase H-digested glycopeptides (0) from MAV-2 gp85; (B) superimposed profiles of ConA-unbound (0) and 10 mm a-methyl mannoside-eluted fractions (0) of MAV-2 gp85 glycopeptides digested with endoglycosidase H; (C) superimposed profiles of the ConAunbound (0) and 10 mm a-methyl mannoside eluted fractions (0) of MAV-1 gp85 glycopeptides digested with endoglycosidase H. methyl mannoside-eluted fraction (fractions 64 through 68, Fig. 5B and C). The small amount of radiolabel eluting between the 14C-labeled (Man)5GlcNAc1 marker and stachyose was recovered in the ConA-unbound fraction, as expected for the N-acetylglucosamine peptides from the endoglycosidase H digestion of hybridand neutral-type glycopeptides.

7 VOL. 45, 1983 DISCUSSION The combined techniques of gel filtration, endo-,-n-acetylglucosaminidase digestion, and ConA-agarose affinity chromatography have been used to analyze the [3H]mannose- and [3H]glucosamine-labeled glycopeptides derived from gp85 of MAV-1 and MAV-2. These studies have demonstrated that the major envelope glycoproteins of both the subgroup A (MAV-1) and subgroup B helper viruses (MAV-2) in the avian myeloblastosis virus complex acquire a diverse array of acidic-type, hybrid-type, and neutraltype asparaginyl oligosaccharides when the viruses are replicated in nontransformed CEF (Table 1). The collections of oligosaccharides from MAV-1 and MAV-2 gp85 were similar to each other and also similar to that acquired by gp85 of a typical subgroup A avian leukosis virus (RAV-1) replicated in equivalent CEF and to the previously described glycopeptides from a nontransforming derivative of PrC RSV (10). The dramatic difference in gel filtration profiles between the [3H]mannose-labeled glycopeptides (Fig. 2 and 4) and the [3H]glucosaminelabeled glycopeptides (Fig. 5) from gp85 of the same viruses was expected because of the much higher ratio of N-acetylglucosamine to mannose in acidic-type versus neutral-type asparaginyl oligosaccharides. If 0-linked, N-acetylgalactosamine-containing oligosaccharides were present in these viral glycoproteins, as recently demonstrated for the gc glycoprotein of herpes simplex type 1 virus (22), these structures should have been labeled with [3H]glucosamine and have appeared in endoglycosidase D- and H-resistant and ConA-unbound glycopeptides of smaller average size than the acidic-type asparaginyl oligosaccharides. The apparent absence of radiolabel in [3H]glucosamine-labeled glycopeptides with these characteristics suggested that MAV-1 and MAV-2 gp85 did not contain 0-linked oligosaccharides, as expected from the reported absence of N-acetylgalactosamine in the major glycoprotein of the standard strain of AMV (glycoprotein encoded by helper virus MAV-1 or MAV-2 or both) (23). In addition to an apparent molecular weight difference between gp85 of MAV-2 and gp85 of MAV-1 and RAV-1, several differences in the glycopeptides were observed: the ratio of ConAunbound glycopeptides (presumably tri- or tetraantennary [or both] acidic type) versus ConAbound glycopeptides (presumably diantennary acidic type) was higher for MAV-1 and RAV-1 gp85 than for MAV-2 gp85, and the ConAunbound glycopeptides exhibited a more heterogeneous size distribution for gp85 from the subgroup A viruses. These studies suggested that the more highly branched acidic-type structures [(NeuNAc ± Gal-GlcNAc)>2(Man)3GlcNAc2- ENVELOPE GLYCOPROTEINS OF AVIAN MAVs 239 (±fucose)-asn] were the predominant class of asparagine-linked oligosaccharides in the glycoproteins from all three viruses. The differences in the oligosaccharide compositions of MAV-1 gp85 and MAV-2 were less than the differences we have previously observed in the glycopeptides of PrC RSV released from transformed versus nontransformed CEF (10). Lai and Duesberg (17) had concluded from earlier studies of subgroup A, C, and D avian retroviruses that strain- or subgroup-specific differences in the size distribution of glycopeptides were minor compared with differences in the same subgroup virus from transformed versus nontransformed cells. More detailed analysis of the envelope glycoproteins of avian retroviruses will be necessary to determine possible differences in the number of oligosaccharide chains per glycoprotein molecule or the distribution of various types of asparagine-linked oligosaccharides along the polypeptide chain for different viral strains. Because of the reported antigenicity of the oligosaccharide moieties of MAV-2 gp85 when the virus is replicated in AMV-transformed myeloblasts (but not from virus replicated in CEF) (27, 28), it will be of particular interest to obtain [3H]mannose- and [ H]glucosamine-labeled virus gp85 from AMV-transformed nonproducer myeloblasts that have been superinfected with MAV-2. The analysis of these radiolabeled glycopeptides in comparison with the present results might reveal compositional or structural alterations in the oligosaccharides that could account for the unusual antigenic properties, such as the possible presence of erythroglycan structures [repeating units of Gal(p1,4)- GlcNAc(,13,3)] in acidic-type asparaginyl oligosaccharides (18, 26). ACKNOWLEDGMENTS We thank Patricia Loh, Bill Lamph, Kevin Christopherson. and David Burnett for technical assistance in part of these studies. This work was supported by Public Health Service grant Al from the National Institute of Allergy and Infectious Diseases and a young investigator research grant from Eli Lilly Research Laboratories to L.A.H. and by the Medical Research Service of the Veterans Administration. LITERATURE CITED 1. Beard, J. W. 1%3. Avian virus growths and their etiologic agents. Adv. Cancer Res. 7: Bonner, W. M., and R. A. Laskey A film detection method of tritium-labeled proteins and nucleic acids in polyacrylamide gels. Eur. J. Biochem. 46: Duesberg, P. H., K. BLster, and C. Moscovid Genetic structure of avian myeloblastosis virus released from transformed myeloblasts as a defective particle. Proc. Natl. Acad. Sci. U.S.A. 77: Duesberg, P. H., G. S. Martin, and P. K. Vogt Glycoprotein components of avian and murine RNA tumor viruses. Virology 41: Etchison, J. R., J. S. Robertson, and D. F. Summers Partial structural analysis of the oligosaccharide moieties

8 240 HUNT AND WRIGHT of the vesicular stomatitis virus glycoprotein by sequential chemical and enzymatic degradation. Virology 78: Galhouse, D. M., and P. H. Duesberg Glycoproteins of avian tumor virus recombinants: evidence for intragenic crossing-over. J. Virol. 25: Hunt, L. A In vitro translation of encephalomyocarditis viral RNA: synthesis of capsid precursor-like polypeptides. Virology 70: Hunt, L. A Sindbis virus glycoproteins acquire unusual neutral oligosaccharides in both normal and lectin-resistant Chinese hamster ovary cell lines. Virology 113: Hunt, L. A Lectin affinity chromatography of Sindbis and Rous sarcoma virus glycopeptides and oligosaccharides. J. Virol. Methods 4: Hunt, L. A., W. Lamph, and S. E. Wright Transformation-dependent alterations in the oligosaccharides of Prague C Rous sarcoma virus glycoproteins. J. Virol. 37: Hunt, L. A., and D. F. Summers Glycosylation of vesicular stomatitis virus glycoprotein in virus-infected HeLa cells. J. Virol. 20: Hunt, L. A., and S. E. Wright Rous sarcoma virus glycoproteins contain hybrid-type oligosaccharides. J. Virol. 39: Hunt, L. A., S. E. Wright, J. R. Etchison, and D. F. Summens Oligosaccharide chains of avian RNA tumor virus glycoproteins contain heterogeneous oligomannosyl cores. J. Virol. 29: Ishlzaki, R., A. J. Langlols, and D. P. Bolognesd Isolation of two subgroup-specific leukemogenic viruses from standard avian myeloblastosis virus. J. Virol. 15: Krantz, M. J., Y. C. Lee, and P. P. Hung Characterization of the major glycoprotein from three strains of Rous sarcoma virus. Arch. Biochem. Biophys. 174: Laemmll, U. K Cleavage of structural proteins during the assembly of the head of bacteriophage T4. Nature (London) 227: Lai, M. M. C., and P. H. Duesberg Differences between the envelope glycoproteins and glycopeptides of avian tumor viruses released from transformed and nontransformed cells. Virology 50: J. VIROL. 18. Li, E., R. Gibson, and S. Kornfeld Structure of an unusual complex-type oligosaccharide isolated from Chinese hamster ovary cells. Arch. Biochem. Biophys. 199: Mardani, D. J., and J. D. Papamatheakds Anomalous behavior of the major avian myeloblastosis virus glycoprotein in the presence of sodium dodecyl sulfate. J. Virol. 26: Mardanl, D. J., and J. D. Papamtheakls Isolation and characterization of the major envelope glycoprotein of avian myeloblastosis virus. J. Biol. Chem. 255: Moscovki, C Leukemic transformation with avian myeloblastosis virus: present status. Curr. Top. Microbiol. Immunol. 71: Olofson, S., S. Jeausson, and E. Lycke Unusual lectin-binding properties of a herpes simplex virus type-i specific glycoprotein. J. Virol. 38: Porter, W. H., and R. J. Winzier Purification and chemical characterization of the major glycoprotein of avian myeloblastosis virus. Arch. Biochem. Biophys. 166: Schmidt, E. V., J. D. Keene, M. Linial, and R. E. Smith Association of 3' terminal RNA sequences with avian leukosis viruses causing a high incidence of osteopetrosis. Virology 116: Smith, R. E., and C. Moscovici The oncogenic effects of nontransforming viruses from avian myeloblastosis virus. Cancer Res. 29: Turco, S. J., J. S. Rush, and R. A. Laine Presence of erythroglycan on human K-562 chronic myelogenous leukemia-derived cells. J. Biol. Chem. 255: Van Eldik, L. J., J. C. Paulson, R. W. Green, and R. E. Smith The influence of carbohydrate on the antigenicity of the envelope glycoprotein of avian myeloblastosis virus and B77 avian sarcoma virus. Virology 86: Van Eldik, L. J., and R. E. Smith Isolation and characterization of the envelope glycoprotein of an avian osteopetrosis virus: effect of host cell on antigenic reactivity. Virology 90: Vogt, P. K A heterogeneity of Rous sarcoma virus revealed by selectively resistant chick embryo cells. Virology 25: