Selective cleavage by endo-beta-N-acetylglucosaminidase H at individual glycosylation sites of Sindbis virion envelope glycoproteins.

Endo-8-N-acetylglucosaminidase H (endo H) was used to probe the relative accessibility of the four asparagine-linked oligosaccharides of El and E2 glycoproteins in intact Sindbis virions. A mutant clone of Chinese hamster ovary cells, clone 15B, which has been shown to lack N-acetylglucosaminyltransferase I, was used as the viral host. In this cell line, high mannose type oligosaccharides are synthesized normally, but formation of complex type glycans is blocked; thus, all N-linked glycans retain endo H-sensitive structures. Using high performance liquid chromatography to separate tryptic glycopeptides from E l or E2, we examined endo H cleavage of oligosaccharides at individual glycosylation sites of intact Sindbis virions. Under nondenaturing conditions, exhaustive endo H digestion resulted in preferential cleavage at glycosylation sites which have complex type glycans in virus grown in wild type Chinese hamster ovary cells (see accompanying paper, Hsieh, P., Rosner, M. R., and Robbins, P. W. (1983) J. Biol. Chem. 258, 2548-2554). Thus, 7480% of the oligosaccharides at a complex site in E2 and 60-63% of the glycans at a complex site in E l were cleaved by endo H, but only 15-168 of the oligosaccharide chains at the high mannose site of E2 were released. At a site in E l which has both high mannose and complex type oligosaccharides in virus from wild type Chinese hamster ovary cells, 23-25% of the oligosaccharides were cleaved by endo H. The selectivity of endo H cleavage was lost when virions were digested with pronase or incubated in detergent. Our findings support the hypothesis that the extent of oligosaccharide processing at glycosylation sites in Sindbis virus glycoproteins is determined primarily by the physical accessibility of oligosaccharides.

Endo-8-N-acetylglucosaminidase H (endo H) was used to probe the relative accessibility of the four asparagine-linked oligosaccharides of El and E2 glycoproteins in intact Sindbis virions. A mutant clone of Chinese hamster ovary cells, clone 15B, which has been shown to lack N-acetylglucosaminyltransferase I, was used as the viral host. In this cell line, high mannose type oligosaccharides are synthesized normally, but formation of complex type glycans is blocked; thus, all N-linked glycans retain endo H-sensitive structures. Using high performance liquid chromatography to separate tryptic glycopeptides from E l or E2, we examined endo H cleavage of oligosaccharides at individual glycosylation sites of intact Sindbis virions. Under nondenaturing conditions, exhaustive endo H digestion resulted in preferential cleavage at glycosylation sites which have complex type glycans in virus grown in wild type Chinese hamster ovary cells (see accompanying paper, Hsieh, P., Rosner, M. R., and Robbins, P. W. (1983) J. Biol. Chem. 258, 2548-2554). Thus, 74-80% of the oligosaccharides at a complex site in E2 and 60-63% of the glycans at a complex site in E l were cleaved by endo H, but only 15-168 of the oligosaccharide chains at the high mannose site of E2 were released. At a site in E l which has both high mannose and complex type oligosaccharides in virus from wild type Chinese hamster ovary cells, 23-25% of the oligosaccharides were cleaved by endo H. The selectivity of endo H cleavage was lost when virions were digested with pronase or incubated in detergent. Our findings support the hypothesis that the extent of oligosaccharide processing at glycosylation sites in Sindbis virus glycoproteins is determined primarily by the physical accessibility of oligosaccharides.
The asparagine-linked oligosaccharides of mammalian glycoproteins have heterogenous structures that generally fall into two categories: "high mannose" and "complex" (reviewed in Refs. 1 and 2). Both classes are synthesized from a common lipid-linked precursor, Glc3Man9GlcNAcn, which is transferred to polypeptides cotranslationally (3,4). The common precursor is extensively modified to yield both complex and high mannose type oligosaccharides. The three terminal glucose residues (5) and possibly a single mannose residue (6) are removed in the endoplasmic reticulum, but subsequent oligosaccharide processing takes place in the Golgi apparatus. Removal of up to four a-1,Z-linked mannose residues by Golgi mannosidases IA and IB (7,8) generates mature high mannose glycans.
Synthesis of complex type oligosaccharides begins with modification of a high mannose precursor, Man5GlcNAcz; the fist step is the addition of an N-acetylglucosamine residue to this protein-linked oligosaccharide (9,10). Mature complex type glycans are produced from GlcNAcMannGlcNAc2 by the removal of 2 more mannose residues and the addition of the peripheral sugars N-acetylglucosamine, galactose, fucose, and sialic acid (reviewed in Ref. 11).
Regulation of oligosaccharide processing at individual glycosylation sites is poorly understood at present. Strain-dependent variations in oligosaccharide composition among closely related strains of influenza virus (12, 13) and murine leukemia virus (14) suggest that the primary amino acid sequence of a glycoprotein can affect the processing of its oligosaccharides. Earlier studies have identified two prerequisites for glycosylation at a specific asparagine residue: location within the sequence -Am-X-Ser(Thr)-(15) and steric accessibility to the oligosaccharide transferase enzyme (16). By analogy, it is possible that a specific amino acid sequence determines whether a particular asparagine residue will carry high mannose or complex oligosaccharides, though no such sequence has ever been identified. An alternative hypothesis is that, while cellular-processing enzymes may not recognize detailed protein structure they will preferentially process oligosaccharide chains which are sterically more accessible. The present study was designed to test this latter possibility.
The enzyme endo H' was used to probe the relative exposure of oligosaccharides of the Sindbis virus envelope glycoproteins E l and E2 in intact virions. Endo H is specific for high mannose oligosaccharides and cleaves between the two proximal GlcNAc residues (17). A mutant clone of CHO cells, clone 15B (18,19), was used as the viral host. These cells lack Golgi N-acetylglucosaminyltransferase I, so processing of oligosaccharides that are normally complex type is halted at the MansGlcNAcz stage, and all N-linked glycans remain sensitive to endo H. Using reverse phase HPLC to separate tryptic glycopeptides containing individual glycosylation sites from E l and E2, we observed that oligosaccharides at sites which have complex oligosaccharides in virus grown in wild type cells were cleaved more readily by endo H than oligosaccharides a t a site which has high mannose oligosaccharides (see accompanying paper (20)). This finding is consistent with the hypothesis that steric accessibility of oligosaccharides is an important factor in the control of oligosaccharide processing. Infection-CHO-K1 cells and clone  15B cells were cultured as monolayers in Alpha Minus minimal  essential medium (DM-325 For infection, confluent monolayers of cells grown in 490-cm' plastic roller bottles were infected with 0.1 plaque-forming unit/cell in 10 ml of medium per roller bottle. Virus was allowed to adsorb to cells for 1 h at 37 "C. Then 9 ml of medium containing 1 mCi D-[~-"H] mannose (20-30 Ci/mmol, New England Nuclear) was added, and the cells were incubated for 18-24 h at 37 "C.

Cell Preparation and Viral
Isolation of mature virions was carried out as described in the accompanying paper (20).
Pronase and Endo H Digestion-Pronase glycopeptides were prepared as described previously (21). Endo H was prepared according to the method of Tarentino and Maley (17). Glycopeptides and intact virions were incubated with endo H (3 pg/ml) for 4 h at 37 "C in 0.1 M Tris-C1, pH 6.8, unless otherwise noted. In control incubations, endo H was incubated without virions in the same buffer. Then endo H activity was assayed by the method of Tarentino and Maley (17), except that 0.1 M Tris-C1, pH 6.8, was used as the buffer.
Separation of Free Oligosaccharides from Glycoprotein-Chloroform/methanol/water extraction was carried out as described previously (21). In an Eppendorf tube, 750 p1 of chloroform/methanol (3:2) and 50 pl of calf serum carrier were added to 100 pl of virions which had been incubated at 37 "C with or without endo H to achieve a final ratio of 3 2 1 chloroform/methanol/water. The mixture was agitated vigorously on a Vortex mixer and centrifuged for 5 min. The organic phase was washed once by the addition of 0.5 ml of water followed by vigorous agitation and centrifugation. The denatured protein was washed twice by sonification in 0.5 ml of water followed by centrifugation. The three aqueous washes were combined with the initial aqueous phase, dried under air, and resuspended in water prior to gel filtration analysis.
Gel Filtration Chromatography-Oligosaccharides and glycopeptides were resolved as described previously on columns (1 X 115 cm) of Bio-Gel P-4 (-400 mesh, Bio-Rad) in 0.1 M pyridinium acetate, pH 6.0 (22). Fractions of 0.5 ml were collected. Bovine serum albumin and ['4C]mannose were added to each sample as markers for the exclusion and inclusion volumes, respectively.
Trypsin Digestion and Reverse Phase HPLC-['HH]Mannose-labeled virions were denatured and then digested with two 30-pg aliquots of trypsin (Worthington, type TRTPCK) for 4 h at 37 "C and analyzed by reverse phase HPLC as described in the accompanying paper (20).

Oligosaccharide Composition of Sindbis Virus Grown in
Wild Type a n d Clone 15B CHO Cells-Sindbis virus grown in ["Hlmannose-labeled wild type or clone 15B CHO cells was treated with pronase and endo H and the oligosaccharides and glycopeptides analyzed by gel filtration chromatography (Table I). In some cases, a single fraction fell between two peaks. The material in these fractions was assigned equally to the two adjacent oligosaccharide species. Such fractions typically contained less than 15% of the total radioactivity associated with either of the adjacent peaks. Virus grown in wild type CHO cells contained endo H-resistant, complex type glycans (66% of the oligosaccharide chains) as well as high mannose oligosaccharides (34%). In contrast, all of the oligosaccharides from virus grown in 15B cells were of the high mannose type, a result consistent with the findings of Gottlieb et al. (18). Identification of high mannose and complex type oligosaccharides based on high resolution gel fiitration chromatography of pronase and endo H-treated virions is consist- CHO or clone 15B cells ["HJMannose-labeled Sindbis virus grown in wild type CHO or clone 15B cells was digested with pronase and endo H and analyzed by gel filtration chromatography as described under "Experimental Procedures." Oligosaccharide species were identified by comparison of their relative elution constants (& values) with those of known standards kindly provided by Dr. S. C. Hubbard (21). The per cent of total chains was calculated by normalizing the recovered radioactivity for the number of mannose residues in each oligosaccharide. CHO Glucose-containing high mannose oligosaccharides have been tentatively identified on the basis of migration on gel fiitration chromatography. Their compositions have not been determined. ent with previously reported findings from a number of laboratories concerning the structures of Sindbis virus oligosaccharides (20,21,(23)(24)(25). While Man5GlcNAcZ constituted only 6% of the total oligosaccharide chains of virus grown in wild type cells, it comprised 59% of the oligosaccharides of virus grown in 15B cells. The almost 10-fold increase in the number of ManaGlcNAc2 oligosaccharides in 15B-grown virus clearly reflected the replacement in these cells of complex type glycans by Man5GlcNAcn. Thus, endo H could be used to probe the accessibility of oligosaccharides in intact Sindbis virions because all N-linked glycans of 15B-grown virus were of the high mannose type.
Analysis of Oligosaccharides Liberated by Endo H Treatment of Intact Virions-We examined the susceptibility of oligosaccharides of virions grown in 15B cells to endo H cleavage under nondenaturing conditions. Incubation of intact [3H]mannose-labeled virions with endo H a t a n enzyme concentration of 3 pg/ml was carried out at 37 "C for 2,4, or 16.5 h.
In order to separate the free oligosaccharides released by endo H from the residual protein-linked glycans, each reaction mixture was extracted with chloroform/methanol/water (3:2:1). The aqueous phase, which contained the free oligosaccharides, was analyzed directly by gel filtration chromatography ( Fig. 1, A, C, and E). Denatured protein, containing residual, covalently linked oligosaccharides, was digested with pronase and endo H, and the resulting oligosaccharides were analyzed by gel filtration chromatography ( Fig. 1, B, D, and

F ) .
Clearly, the susceptibility of the 15B-grown virion oligosaccharides to endo H was not random; virtually all of the oligosaccharide chains released under these nondenaturing conditions were either MansGlcNAc (90-95%) or ManlGlcNAc (4-6%) (Fig. 1, A, C, and D). Gel  peptides from ["Hlmannose-labeled Sindbis virus grown in clone 15B cells were resolved by HPLC, the same four glycosylation sites were observed (Fig. 2).
To determine the oligosaccharide composition at each glycosylation site in Sindbis virus grown in 15B cells, HPLC fractions corresponding to a single glycosylation site were pooled. In repeated trials, peaks I and I1 were well resolved and separated by a fraction which contained no greater than 12% of the total counts/min in either peak alone. Peaks IV and V were separated by a single fraction which contained glycopeptides from both sites. Material from this fraction was analyzed separately from the other peak fractions which were free of contaminating glycopeptides. The [:'H]mannose-labeled glycopeptides were treated with pronase and endo H and analyzed by gel filtration chromatography. The results of such an analysis, shown in Table 11, are consistent with the replacement of the complex type oligosaccharides found in wild type CHO cells by ManBGlcNAca in clone 15B cells. In virus grown in wild type CHO cells, E l has complex type oligosaccharides at one site (HPLC peak V) and both complex and high mannose type oligosaccharides at a second site (peak 111) with the majority of high mannose oligosaccharides being Man5GlcNAca; E2 has high mannose oligosaccharides at one site (peaks I+IV) and complex oligosaccharides at the second site (peak 11). When the same virus is grown in 15B cells, both glycosylation sites of E l (HPLC peaks I11 and V) have predominantly Man5GlcNAc:! oligosaccharides; in E2, the high mannose site (peaks I+IV) has larger high mannose structures in addition to MansGlcNAc2, while at the complex type site (peak 11), MansGlcNAc2 oligosaccharides predominate. A small amount of material at three of the four sites (peaks I+IV, 11, and V) has been tentatively identified on the basis of migration in gel filtration chromatography, as large, incompletely processed high mannose oligosaccharides containing terminal glucose residues. As was noted for virus grown in wild type CHO cells, three of the four glycosylation sites (HPLC peaks I+IV, 11, and V) are glycosylated at roughly the same frequency in 15B-grown virus; the remaining site (peak 111) is glycosylated about 80% as often as the other three sites.
HPLC Analysis of Selective Endo H Cleavage-Having Oligosaccharide compositions a t individual glycosylation sites of Sindbis virus grown in 15B cells ['HIMannose-labeled tryptic peptides from virions grown in 15B cells were separated by reverse phase HPLC as described in Fig. 2. Fractions containing radiolabeled glycopeptides from a single HPLC peak were pooled, desalted on Sephadex G-10-and lyophilized. The resulting material was treated with pronase and endo H and analyzed by gel filtration chromatography as described in Fig. 1. The per cent of total oligosaccharides was determined for each glycosylation site by normalizing for the number of mannose residues in each chain.

Composition of oligosaccharides released by endo H
In two independent experiments, ['Hlmannose-labeled Sindbis virions grown in clone 15B cells were incubated under nondenaturing conditions in 0.1 M Tris-C1, pH 6.8, for 4 h at 37 "C in the presence of 3 pg/ml of endo H. After extraction with chloroform/methanol/water (3:2:1), free oligosaccharides released by endo H were analyzed by gel filtration chromatography as described in Fig. 1. Denatured protein containing residual, covalently linked glycans was digested with trypsin and analyzed by HPLC (Table IV). Oligosaccharides resolved by gel filtration were identified by comparison of their relative elution constants (& values) with those of known standards. The per cent of total released oligosaccharides was determined by normalizing for the number of mannose residues in each chain.  [JH]Mannose-labeled virions grown in 15B cells were incubated with or without endo H under nondenaturing conditions followed by extraction as described in Fig. 1. Gel filtration analysis of the released oligosaccharides from two independent trials are shown in Table 111. Denatured protein containing covalently linked residual glycans (+endo H) or total virion oligosaccharides (-endo H) was digested with trypsin and the tryptic peptides resolved by HPLC as described in Fig. 2 The per cent of total oligosaccharides released was calculated for eaeh glycosylation site as follows. The fraction of endo H-released ManeGlcNAcl was calculated for each site. This fraction was compared to the percentage of total chains attributable to MansGlcNAcl for each site (Table 11) to arrive at the percentage of total oligosaccharides released. For purposes of this table, it was assumed that all endo H-released 3H counts/min were attributable to Man5GlcNAcl. For example, for HPLC peak V, trial 1, the initial total radioactivity in peak V is 4,570 cpm (-endo H). From Table 11: 78% of radioactivity in peak V is associated with Man5GlcNAcl. 0.78 X 4,570 cpm = 3,590 cpm of Man5GlcNAcl initially.
From above: 2,540 cpm of MansGlcNAcl were released by endo H. 2,540 cpm released/3,590 cpm initial = 71% of ManeGlcNAcl chains released by endo H at peak V. From Table 11: 85% of total chains at peak V are MansGlcNAcl. Therefore, 0.71 x 85% = 60% of total chains at peak V were released by endo H. by guest on March 24, 2020 http://www.jbc.org/ Downloaded from characterized the oligosaccharide composition at individual glycosylation sites of Sindbis virus grown in 15B cells, we examined the availability of these oligosaccharides to endo H under nondenaturing conditions. In two independent experiments, [3H]mannose-labeled virions from 15B cells were incubated under nondenaturing conditions with or without endo H for 4 h a t 37 "C. After chloroform/methanol/water extraction, the aqueous phase containing free oligosaccharides from the endo H-treated sample was analyzed directly by gel filtration chromatography (Table  111). The denatured proteins from the endo H-treated and control (-endo H ) samples were treated with trypsin and analyzed separately by reverse phase HPLC (Table IV).
The results of this analysis confirmed that oligosaccharides at complex type glycosylation sites were preferentially released by endo H (Table IV). Thus, 7440% of the oligosaccharides at the complex site in E2 (peak 11) and 6043% of the oligosaccharides at a complex site in E l (peak V) were cleaved by endo H, but only 15-16% of the oligosaccharide chains at the high mannose site of E2 (HPLC peaks I+IV) were released. At a site in E l which has been high mannose and complex type oligosaccharides in virus grown in wild type CHO cells (peak 111), 23-25% of the oligosaccharides were cleaved.
Analysis of Oligosaccharides Liberated by Endo H in the Presence of Detergent-To explore the possible role of tertiary polypeptide conformation or glycoprotein-lipid interactions in determining the accessibility of oligosaccharides, we incubated virions with endo H in the presence of ionic or nonionic detergents.
[3H]Mannose-labeled virions grown in 15B cells were heated a t 100 "C for 2 min in 0.2% SDS and 1% pmercaptoethanol or suspended in 0.2% NP-40 at room temperature prior to addition of endo H. Detergent was present during the incubation with endo H, which was performed as described in the legend to Fig. 1. A control sample was incubated for 4 h a t 37 "C without detergent or enzyme. The endo H-released oligosaccharides from each of the two detergent-treated samples were analyzed by gel filtration chromatography (Fig. 3, A and C). The denatured protein, containing unreacted oligosaccharides in the SDS and NP-40 samples or total virion oligosaccharides in the control sample, was treated with pronase and endo H and analyzed by gel fiitration chromatography (Fig. 3, B, D, and E ).
The composition of oligosaccharides released by endo H was identical for virions incubated in SDS or NP-40 (Fig. 3, A and C). Comparison of endo H-released and -resistant oligosaccharides in the two detergent-treated samples revealed that over 95% of total Man5GlcNAcz chains were cleaved in the presence of detergent as compared to only 65% of the total ManaGlcNAcz chains released by endo H under nondenaturing conditions (Fig. 1C). Furthermore, while less than 5% of high mannose oligosaccharides containing 6 to 9 mannose residues were cleaved under nondenaturing conditions (Fig. ID), virtually all the ManGGlcNAc? and Man7GlcNAcp, two-fifths of the MansGlcNAc?, and one-fifth of the Man9GlcNAcz oligosaccharides were cleaved by endo H in the presence of detergent (Fig. 3, B and D). Approximately 20% of the total oligosaccharide chains remained resistant to cleavage by endo H. AS the gel filtration data in Fig. 3 show, the composition of uncleaved chains is biased toward MansGlcNAcz and Man9GlcNAcn as well as oligosaccharides which are tentatively identified as glucose-containing high mannose structures. Since endo H retained over 90% of its initial activity following incubation of the enzyme in detergent under conditions used above, the incomplete release of oligosaccharides was not due to enzyme inactivation, and is probably attributable to incomplete denaturation of E l and E2 glycoproteins. Reverse phase HPLC analysis of tryptic glycopeptides from the same experiment shown in Fig. 3, containing residual oligosaccharides, revealed that endo H released oligosaccharides from all four glycosylation sites to approximately the same extent when virions were treated with detergent prior to incubation with endo H (data not shown). Approximately 80% of the oligosaccharide chains from HPLC peaks I1 and V, 80-85% of the chains a t peak I11 and 70-75% of the chains a t peaks I+IV were released by endo H in the presence of either SDS or NP-40. The gel filtration chromatography and HPLC data suggest that a subset of virion oligosaccharides present by guest on March 24, 2020 http://www.jbc.org/ Downloaded from at all four glycosylation sites and consisting almost exclusively of large, high mannose glycans are inaccessible to endo H under our conditions of detergent treatment.

DISCUSSION
Endo H was used to probe the relative accessibility of the four asparaginyl glycosylation sites of intact Sindbis virions. Clone 15B CHO cells were chosen as the host, since they are incapable of converting protein-linked Man5GlcNAc2 to complex type glycans. Thus, sites which carry exclusively complex type oligosaccharides ("complex sites") in virus grown in wild type CHO cells have Man4-5G1cNAc2 in virus from clone 15B cells; the oligosaccharide composition at a site which carries high mannose glycans ("high mannose site") is similar in virus from both hosts. Our analysis revealed that virtually all of the oligosaccharide chains released by endo H under nondenaturing conditions from intact clone 15B-grown Sindbis virions were either MansGlcNAc (90-93%) or Man4GlcNAc (4-696).
Furthermore, the relative amount of endo H cleavage at each of the four glycosylation sites of the viral glycoproteins correlated with the extent of oligosaccharide processing observed for virus grown in wild type cells (20). Thus, only 15-16% of the oligosaccharides located at a high mannose site in E2 were cleaved by endo H, while 74-80% of the oligosaccharides at a complex site in E2 and 60-63% of the oligosaccharides a t a complex site in E l were cleaved. Endo H cleaved an intermediate number of oligosaccharide chains (23-25%) from the remaining E l glycosylation site, which carries both high mannose and complex glycans in virus from wild type CHO cells. Interestingly, this "intermediate" site exhibits marked hostdependent variation of oligosaccharide composition; in virus grown in chicken embryo fibroblasts it carries exclusively high mannose glycans, while in virus grown in baby hamster kidney cells it carries predominantly complex glycans (20). Preferential cleavage by endo H a t complex glycosylation sites was abolished when the native conformations of E l and E2 glycoproteins were disrupted.
Thus, after digestion of virions with pronase, endo H released all oligosaccharides present in the glycopeptides. Likewise, when virions were heated at 100 "C in 0.2% SDS and P-mercaptoethanol prior to endo H digestion, approximately 80% of the oligosaccharides at all four sites were cleaved. The residual oligosaccharides were almost exclusively larger high mannose structures and probably remained inaccessible to the enzyme due to incomplete denaturation of the glycoproteins under our conditions. These results suggest that the tertiary conformation of the polypeptide may determine the accessibility of oligosaccharides in Sindbis virus glycoproteins. However, preferential cleavage at complex sites by endo H was also abolished when virions were incubated with the nonionic detergent NP-40. It is possible, therefore, that glycoprotein-lipid interactions affect the accessibility to endo H of oligosaccharides of E l and E2. Alternatively, the conditions of NP-40 incubation used above may have resulted in disruption or alteration of the native conformations of these glycoproteins. Our results strongly support the hypothesis that the extent of oligosaccharide processing at individual glycosylation sites is determined primarily by the physical accessibility of oligosaccharides. Recent findings from other laboratories are also consistent with this steric model. Natowicz et al. (26) and Howard et al. (27) observed that phosphorylated oligosaccharides and smaller neutral glycans of native P-glucuronidase, a lysosomal enzyme, were released by endo H while larger, nonphosphorylated chains were resistant to cleavage. Similarly, Trimble et al. (28) found that in yeast, more extensively processed oligosaccharides of nondenatured invertase and carboxypeptidase Y were more susceptible to endo H than less processed oligosaccharides.
It should be noted that while we have clearly established a correlation between the extent of endo H cleavage of oligosaccharides in intact Sindbis virions and the extend of oligosaccharide processing which occurs at individual glycosylation sites, we do not know to what degree the conformations of E l and E2 in mature virions resemble those of the intracellular processing intermediates. Rice and Strauss (29) demonstrated t h a t E l and E2 reside on the surface of mature virions in repeating hexamer units stabilized by strong, noncovalent interactions between E l and E2. The same study also showed that El and pE2, a precursor to E2, can form weekly stabilized heterodimers intracellularly (29). Thus, it is possible that the tertiary conformations of E l and E2 present in mature virions are very similar to the conformations which are recognized by oligosaccharide-processing enzymes in the Golgi apparatus.
Although our results emphasize the importance of steric accessibility of oligosaccharides to cellular processing enzymes, other factors may also affect oligosaccharide processing. In some cases, recognition of specific polypeptide conformation by a processing enzyme may direct the synthesis of unusual glycans with structures falling outside the simple high mannose and complex categories discussed here. For example, the Golgi enzyme which transfers N-acetylglucosamine-lphosphate to the high mannose glycans of lysosomal enzymes appears to recognize a structural determinant(s) unique to these glycoproteins (30). It is possible that specific recognition also occurs in the synthesis of other specialized oligosaccharides, such as the large polylactosamine glycans present on a subset of cell surface proteins in erythroid cells (31, 32).
We are currently pursuing approaches to identify specific factors which influence oligosaccharide processing and to examine the relationship between oligosaccharide accessibility and processing in a variety of glycoproteins.