Glycosylation Requirements for Intracellular ... - Journal of Virology

Vol. 66, No. 12

JOURNAL OF VIROLOGY, Dec. 1992, p. 7136-7145

0022-538X/92/127136-10$02.00/0 Copyright © 1992, American Society for Microbiology

Glycosylation Requirements for Intracellular Transport and Function of the Hemagglutinin of Influenza Virus PATRICIA J.

GALLAGHER,1 JEAN M. HENNEBERRY,2 JOSEPH F. SAMBROOK,2

AND MARY-JANE H. GETHING3* Departments of Physiology' and Biochemistry,2 and Howard Hughes Medical Institute,3 University of Texas Southwestern Medical Center at Dallas, 5323 Harry Hines Boulevard, Dallas, Texas 75235

Received 15 June 1992/Accepted 21 August 1992

The contribution of each of the seven asparagine-linked oligosaccharide side chains on the hemagglutinin of the A/Aichi/68 (X31) strain of influenza virus was assessed with respect to its effect on the folding, intracellular transport, and biological activities of the molecule. Twenty mutant influenza virus hemagglutinins were constructed and expressed, each of which had one or more of the seven glycosylation sites removed. Investigations of these mutant hemagglutinins indicated that (i) no individual oligosaccharide side chain is necessary or sufficient for the folding, intracellular transport, or function of the molecule, (ii) at least five oligosaccharide side chains are required for the X31 hemagglutinin molecule to move along the exocytic pathway to the plasma membrane, and (iii) mutant hemagglutinins having less than five oligosaccharide side chains form intracellular aggregates and are retained in the endoplasmic reticulum.

niently studied (36). In other cases, underglycosylated polypeptides are partially or severely affected, displaying defects in folding and transport that may or may not depend on the position in the molecule of the attachment site of the missing side chain (22, 29, 31, 32, 34, 40, 49, 55). This type of study can be most informative when performed on a glycoprotein whose three-dimensional structure is available, facilitating the appraisal of the potential role of individual side chains in the folding or stabilization of the molecule. The hemagglutinin (HA) glycoprotein of influenza virus is a well-characterized integral membrane protein whose pathway of folding and final three-dimensional structure have been studied in detail (3, 4, 15, 18, 41, 42, 53). Wild-type HA oligomerizes in the ER to form a homotrimeric structure and is then transported along the secretory pathway to its final destination on the plasma membrane (3, 4, 15, 57). Misfolded HA molecules are not transported to the plasma membrane but are retained within the ER in complexes associated with the ER chaperone binding protein (BiP) (15, 25). Determination of the structure of the HA molecule from the A/Aichi/68 (X31) viral strain revealed the locations of the seven N-linked oligosaccharide side chains on each of the subunits in the trimeric structure (50, 56). As shown in Fig. 1, two of the side chains are attached to asparagine residues (Asn-81 and Asn-165) located on the surface of the globular domain of the molecule, while four side chains (attached to Asn-8, Asn-22, Asn-38, and Asn-483) are located on the stem domain. The seventh site is located near the flexible hinge region of the molecule that separates the globular and stem domains (56). All seven glycosylation consensus sites on X31 HA are utilized, and the majority of the side chains on the mature molecule are of the complex type, being trimmed and then terminally modified with galactose, fucose, and sialic acid residues in the Golgi apparatus (9, 56). However, the oligosaccharide side chains attached to Asn-165 (on the top of the globular domain) and Asn-285 (in the hinge region) are not processed in the Golgi apparatus and remain of the mannose-rich type containing only N-acetylglucosamine and mannose residues (56). This lack of processing is probably due to the inaccessibility to glycosidase and glycosyltransferase enzymes of these side chains after trimerization of the

Many cell surface and secreted proteins are modified posttranslationally by the covalent attachment of asparagine-linked (N-linked) oligosaccharide moieties (reviewed in reference 26). The initial addition of the carbohydrate side chains occurs as nascent secretory and transmembrane proteins are translocated into the endoplasmic reticulum (ER), where mannose-rich oligosaccharides having the composition Glc3Man9GlcNAc2 are transferred from a dolichol PPi donor to asparagine residues presented in the consensus sequence Asn-X-Ser/Thr (where X is any amino acid except proline and perhaps aspartic acid) (24, 26). Processing of these large oligosaccharides begins in the ER and continues as the glycoproteins move to and through the Golgi apparatus, generating two major classes of side chains: mannoserich chains containing N-acetylglucosamine and variable amounts of mannose, and complex chains containing N-acetylglucosamine, mannose, galactose, fucose, and sialic acid (26). A variety of functions for N-linked oligosaccharides have been suggested (36, 39), including (i) influencing the folding of the polypeptide chain, (ii) maintenance of protein solubility, (iii) targeting of glycoproteins to various subcellular compartments and to the cell surface, (iv) protection of the polypeptide backbone against proteolytic degradation, (v) modulation of biological activity, and (vi) promotion of specific adhesion of cells during development. Initial studies of the role of modification with N-linked oligosaccharides utilized drugs such as tunicamycin (11) to block the addition of all N-linked side chains to all usually glycosylated cellular proteins (reviewed in reference 36). In recent years, sitedirected mutagenesis of individual glycosylation consensus sequences has been used to assess the role of individual N-linked side chains on a glycoprotein synthesized in an otherwise normal cell (12, 29-32, 34, 45, 49, 54). Both types of studies have revealed great variability in the consequences of removal of one or more N-linked side chains. In some cases, total lack of glycosylation has no apparent effect on the protein, at least for those attributes (folding, intracellular transport, function, and stability) that are conve*

Corresponding author. 7136

VOL. 66, 1992

GLYCOSYLATION OF INFLUENZA VIRUS HEMAGGLUTININ

NXAM XHA

As. 8,22,38,81,165,285,483

sns,RESE2

COMMENT

8-

As -,22,38,81,165,285,483

single site missing

22-

Asn 8, -,38,81 ,165,285,483

38-

Asn 8,22, -,81,165,285,483

GLOBULAR DOMAIN

81-

Ass 8,22,38, -,165,285,483

165-

As. 8,22,38,81, -,285,483

285-

As. 8,22,38,81,165, -,483

483-

As. 8,22,38,81,165,285, -

N

STEM

DOMAIN

7137

wild-type

unglycosylated HA

8+

Asn8

22+

Asn22

38+

Asn38

81+

Asn81

single site present

165+

AsuI65

285+

Asn285

483+

Asn483

CHO-3+

Asn8l,165,285

three sites on globular head and hinge

CHO-4+

Asn8,22,38,483

four sites on stem

CHO-5A+

Asn8,38,81,165,285

five sites

CHO-5B+

Asn8,22,38,81 ,483

five sites, both mannose-rich sites at 165 and 285 missing

CHO-SC+

Asn8,22,38,285,483

five sites, two globular domain missing

i FIG. 1. HA mutants lacking glycosylation sites and schematic illustration of the locations of glycosylation sites on the ectodomain of an HA monomer. Schematic drawing of the ectodomain of a monomer of the A/Aichi/2/68 (X31) HA showing the locations of the seven glycosylation consensus sites and the stem and globular domains (adapted from reference 55). Numbers indicate the asparagine residue to which the side chain is covalently linked. The table lists the sites present in each of the glycosylation mutants.

molecule in the ER. These two oligosaccharides are the only whose mobility in the protein crystal is restricted sufficiently to allow definition of the positions of all the sugar residues in the side chain; one nestles in the notch formed between the tops of the globular domains, while the other is tucked into a cleft formed by the hinge region (56). In this study, we used oligonucleotide-directed mutagenesis to investigate the contribution of the individual carbohydrate side chains on X31 HA. Our results show that of the seven oligosaccharides on this HA molecule, no individual side chain is sufficient or essential for folding and transport to the cell membrane. Further, a minimum of five N-linked glycosylation sites are required for folding, trimerization, and intracellular transport. ones

MATERIALS AND METHODS

Recombinant DNA techniques. Buffers and reaction conditions for restriction enzymes, T4 DNA ligase, and DNA polymerase I were those listed by the commercial source (New England Biolabs, Beverly, Mass.). Isolation of DNA fragments, preparation of plasmid DNAs, and other standard

recombinant DNA techniques were done as described previously (43). Site-specific mutagenesis of HA cDNA. Construction of M13mpl9 containing the full-length cDNA encoding the X31 HA protein has been described previously (9, 12). Seven 18-mer oligonucleotides synthesized to be complementary, except for the required mismatches, to HA cDNA sequences encoding the seven glycosylation consensus sites (Asn-XSer/Thr) were utilized to substitute codons for alanine or glycine residues in place of those encoding serine or threonine residues in the seven consensus sequences present in X31 HA (Table 1). To obtain mutant cDNAs lacking a single natural consensus site, we performed oligonucleotide-directed mutagenesis (58) using a single oligonucleotide designed to produce the desired change. Mutant cDNAs having no sites (N), only a single natural glycosylation site, or a combination of sites were generated by performing successive site-directed mutagenesis procedures on previously mutagenized templates. For each mutant HA cDNA, the desired nucleotide substitutions were verified by DNA sequencing (44). Expression of the mutant HA proteins in CV-1 cells was facilitated by

7138

J. VIROL.

GALLAGHER ET AL.

TABLE 1. Nucleotide substitutions and corresponding amino acid changes in mutants with alterations in natural glycosylation consensus sequences Name

8(-)

22(-)

38(-) 81(-) 165(-)

285(-)

483(-)

Sequence of wild type

Asn-Ser-Thr AAC-AGC-ACA Asn-Gly-Thr AAC-GGA-ACA Asn-Ala-Thr AAT-GCT-ACT Asn-Glu-Thr AAT-GAG-ACA Asn-Val-Thr AAC-GTG-ACT Asn-Gly-Ser AAT-GGA-AGT Asn-Gly-Thr AAT-GGG-ACT

42°C for 30 min. These temperatures were accurately maintained throughout the labeling and chase periods until the cell extracts were prepared.

Sequence of mutant

Asn-Ser-Ala

AAC-AGC--jCA

Asn-Gly-Ala AAC-GGA-GCA Asn-Ala-Ala

AAT-GCT-GjCT

Asn-Glu-Ala AAT-GAG-GCA Asn-Val-Ala AAC-GTG-GCT Asn-Gly-Gly

AAT-GGA-GjGT Asn-Gly-Ala AAT-GGG-GCT

subcloning the 1,608-bp ClaI-BamHI DNA fragment encoding the HA ectodomain into the SVEXHA viral expression vector (9). Generation of SV40-HA recombinant virus stocks and infection of CV-1 cells. Simian virus 40 (SV40)-HA recombinant genomes containing wild-type and mutant forms of the X31 HA gene were transfected into CV-1 cells by using DEAE-dextran and chloroquine as previously described (8). High-titer virus stocks were developed and used to infect fresh monolayers of CV-1 cells for analysis of the biosynthesis and intracellular transport of wild-type and mutant HA proteins. Expression of the HA from the A/Japan/305/57 influenza virus strain was achieved by using the SVEHA3 vector described previously (16). Conditions for growth and infection of CV-1 cells were as described previously (8). Analysis of intracellular transport and biological activities of wild-type and mutant HAs. The intracellular transport of wild-type and mutant HA proteins was analyzed by pulsechase radiolabeling of infected cells with [35S]methionine (12, 15). HA displayed on the cell surface was detected by the addition of trypsin (10 ,ug/ml) to the medium above the intact cell monolayer for the last 15 min of the chase (9). When we desired to block N-linked glycosylation of HA molecules, tunicamycin (5 ,ug/ml) was added to the medium for 2 h before labeling and throughout the labeling procedure (12, 15). Cell extracts were immunoprecipitated with polyclonal anti-X31 or anti-Japan HA rabbit antiserum and then resuspended in sodium dodecyl sulfate-polyacrylamide gel

electrophoresis (SDS-PAGE) sample buffer (50 mm Tris [pH 6.8], 0.1 M dithiothreitol [DTT], 2% SDS, 0.1% bromphenol blue, 10% glycerol) before analysis by SDS-PAGE as described previously (15). To detect aggregates, fully dissociated monomers, partially dissociated dimers, and undissociated trimers of HA, immunoprecipitates of HA from cell extracts were resuspended before electrophoresis in SDSPAGE sample buffer that did not contain 0.1 M DTJ. Cell fusion and erythrocyte binding assays were performed as described previously (9, 12). Digestion with endo-1-Nacetylglucosaminidase H (endo H) was done as described previously (37). Finally, the biosynthesis and transport of the wild-type and mutant proteins at 30 and 42°C was analyzed as described above for 37°C with the following modifications. CV-1 cells were infected with recombinant viruses and incubated at 37°C for 36 to 40 h. Before being labeled, the infected cells were preincubated at either 30 or

RESULTS HAs from different strains of influenza virus differ in their requirement for N-linked glycosylation. The HA molecules encoded by the A/Japan/305/57 (H2N2) strain of influenza virus and the X31 recombinant of the A/Aichi/68 (H3N2) strain display only 42% identity in amino acid sequence (13) but almost certainly have very similar three-dimensional structures (56). These two HA molecules also differ in the number of N-linked oligosaccharides attached to the polypeptide chain. As discussed above, X31 HA has seven oligosaccharides, of which only five are processed to the complex type (9, 33, 56). Japan HA has five side chains (8, 51), all of the complex type (8, 33). Four of these side chains are attached to asparagine residues in the stem domain at positions identical or very close to those occupied in X31 HA by asparagines 22, 38, 285, and 483 (13) (Fig. 1). The fifth site is located at the top of the globular domain in a position corresponding to Asn-165 in the X31 HA molecule (Fig. 1). The two attachment sites on X31 HA that have no counterpart in the Japan HA are at Asn-8 at the base of the stem domain and Asn-81 on the globular domain. To compare the glycosylation requirements of these closely related proteins, we analyzed the transport properties of Japan and X31 HA polypeptides synthesized in the presence of tunicamycin. Cells infected with SVEHA3 or SVEXHA virus stocks were pulse-labeled with [35S]methionine for 5 min at 37°C and then chased in medium containing nonradioactive methionine for 3 h. During the last 15 min of chase, the intact cell monolayers were treated or not with medium containing trypsin (5 ,ug/ml), which cleaves any HA molecules exposed on the cell surface into their component HA1 and HA2 subunits (9). Cell extracts were then prepared, and the HA molecules were immunoprecipitated and analyzed by SDS-PAGE. As described previously, the majority of wild-type, glycosylated Japan or X31 HA molecules (gHAO) are transported to the cell surface, where they become accessible to cleavage into gHA, and gHA2 subunits by trypsin in the medium (Fig. 2). Nonglycosylated Japan HA polypeptides synthesized in the presence of tunicamycin (tHAO) were also efficiently transported to the plasma membrane during the 3-h chase since treatment of the intact cells with trypsin resulted in cleavage of the majority of the polypeptides into nonglycosylated tHA, and tHA2 subunits. By contrast, as reported previously (25), X31 HA molecules were not transported to the plasma membrane when the addition of N-linked carbohydrate side chains was inhibited by tunicamycin. Nonglycosylated X31 tHAO molecules were not cleaved by exogenous trypsin during the 3-h chase (Fig. 2). Nor did they become accessible to the protease when the chase was extended to 24 h (data not shown). Thus, Japan and X31 HA molecules differ in their requirements for glycosylation for their transport to the plasma membrane. This observation raises the possibility that one or more of the individual oligosaccharide side chains on X31 HA, perhaps those lacking on Japan HA, might be essential for the folding and intracellular transport of the molecule. Construction and expression of mutant X31 HA molecules lacking natural glycosylation sites. To evaluate the importance of individual N-linked oligosaccharides in the folding and transport of X31 HA, we constructed cDNAs encoding 20 mutant HA proteins lacking either single glycosylation

GLYCOSYLATION OF INFLUENZA VIRUS HEMAGGLUTININ

VOL. 66, 1992

JAPAN HA Tunic Trypsin

-gHA0

a

68K- *

O

a

-

H 1

-gHA2

92K

.1

I.i

-.AmIL

_

_

_

_

-

M

or,o0er'

46K-

-tHA1

-tHA2

n - --;Dinle X~~~ _~ ~u _-~ 02

-gHA0

68K-

-tHAo

-gHA1

5

_ 30K-

2h chase -DTT

+DTT

- + - + - - + +

I*

46K-

2h chase

X31 HA

7139

30K -

gHA2

FIG. 2. Expression of glycosylated and nonglycosylated forms of HA proteins from the A/Japan/30/57 and A/Aichi/2/68 strains of influenza virus. CV-1 cells infected with recombinant viruses expressing either Japan or X31 HA protein were labeled for 5 min with [35S]methionine and then incubated in medium containing an excess of nonradioactive methionine for 3 h in the presence or absence of tunicamycin (Tunic). During the last 15 min of the chase period, trypsin (5 pLg/ml) was added to the medium above the intact monolayer. Cell lysates were immunoprecipitated with polyclonal antibodies specific for either the Japan or the X31 HA molecule. gHAO, gHA1, and gHA2 represent the glycosylated forms of the full-length HA polypeptide and the HA1 and HA2 subunits, respectively; tHAO, tHA1, and tHA2 represent the nonglycosylated forms of these polypeptides synthesized in the presence of tunicamycin. Numbers on left show molecular mass in kilodaltons.

sites or combinations of sites using oligonucleotide-directed mutagenesis (Fig. 1). Mutant HA proteins of the first group [8(-), 22(-), 38(-), 81(-), 165(-), 285(-), 483(-)] lack a single glycosylation site but retain the six other natural sites. A second group of mutant HA proteins either have all seven consensus sites ablated (N) or have only a single glycosylation site [8(+), 22(+), 38(+), 81(+), 165(+), 285(+), 483(+)]. The third group retains specific combinations of three, four, or five glycosylation sites [CHO(+3), CHO(+4), CHO(+5)A, CHO(+5)B, CHO(+5)C]. At six of the seven glycosylation sites, oligonucleotide-directed mutagenesis of X31 HA cDNA was used to substitute an alanine residue for the serine residue in the consensus sequence (Table 1). This substitution of a small, neutral amino acid is unlikely to significantly disrupt the conformation of the folded protein, because the seven naturally occurring consensus sites in the X31 HA polypeptide occupy exposed, surface positions in the quaternary structure of the native molecule (56). In mutant 285(-), the desired change introduces a glycine residue in place of serine. Although glycine is an amino acid noted for increasing main-chain flexibility (46), its substitution at residue 287 occurs within a small disulfide-bonded loop and would not be expected to significantly affect the folding of the whole molecule. To analyze the phenotypes of the mutant glycoproteins, we constructed SV40-HA recombinant genomes in which either the wild-type X31 or the 20 mutated cDNAs replaced the coding sequences of the late region of SV40 DNA (see Materials and Methods). These recombinant genomes were transfected into CV-1 cells, and high-titer recombinant virus stocks were developed. After infection of fresh monolayers

FIG. 3. Detection of trimeric structures formed by HA mutants lacking one glycosylation site. CV-1 cells infected with recombinant viruses expressing wild-type or mutant X31 HA proteins were labeled for 5 min with [35S]methionine and then incubated in medium containing an excess of nonradioactive methionine for 2 h at 37'C. Cell extracts were prepared and immunoprecipitated with a polyclonal anti-X31 HA antiserum. Immunoprecipitates were analyzed by SDS-PAGE under nonreducing conditions (-DTT) or after reduction (+DTT). Numbers on left show molecular mass in kilodaltons.

of CV-1 cells with these virus stocks, a number of different assays were performed to analyze the phenotypes of the wild-type and mutant HA proteins. The state of assembly or aggregation of the polypeptides was determined by analysis by SDS-PAGE under reducing and nonreducing conditions (47). The time course and efficiency of movement of the proteins from the ER to the Golgi apparatus was measured by assaying the rate and extent of acquisition of resistance to cleavage of oligosaccharide chains by endo H (26), while the arrival of HA molecules at the cell surface was monitored by assaying their accessibility to proteolytic cleavage into HA1 and HA2 polypeptides by trypsin added to the external medium (9). The biological activities of the mutant proteins were assessed by erythrocyte binding and cell-cell fusion assays (16, 52). The details of these procedures are given in Materials and Methods. Mutant X31 HAs lacking any single, natural glycosylation site assemble into trimers and are efficiently transported to the plasma membrane. Cells expressing wild-type X31 HA or mutant proteins lacking a single glycosylation site were pulse-labeled with [35S]methionine for 5 min at 37°C and then chased in medium containing nonradioactive methionine for 2 h. Cell extracts were then prepared, and the HA molecules were immunoprecipitated and analyzed by SDS-PAGE. Under reducing conditions (plus DTT), the seven mutant proteins each migrated as a single major band whose mobility was slightly faster than that of the wild-type protein (Fig. 3, left panel). This mobility change, which is consistent with the loss of a single oligosaccharide side chain, was more apparent after cleavage of the polypeptides into HA1 and HA2 subunits (Fig. 4, see below). When the wild-type and mutant HA polypeptides were synthesized in the presence of tunicamycin, all migrated with identical molecular masses on SDS-PAGE (data not shown), indicating that the differences in mobility of the HA species seen in Fig. 3 and 4 reflect changes in the complement of oligosaccharide side chains rather than alterations to the polypeptide backbone. To determine the oligomerization state of the wild-type and mutant HAs, we also analyzed the immunoprecipitated proteins by SDS-PAGE under nonreducing (no DTT) conditions (Fig. 3, right panel). As described previously (6), HA trimers are not completely dissociated into their component monomers by heating in SDS unless a reducing agent is

7140

J. VIROL.

GALLAGHER ET AL.

5'

37°C

pulse

2h

present, so that intact trimers, partially dissociated dimers, and fully dissociated monomers are all observed after electrophoretic separation under nonreducing conditions. The autoradiograph presented in Fig. 3, right panel, shows that the pattern of migration of all the mutant proteins is identical to that of the wild-type X31 HA, indicating that they all assemble into trimeric molecules that can be partially dissociated by SDS, yielding the same ratios of monomers, dimers, and intact trimers. These results demonstrate not only that the loss of a carbohydrate side chain is tolerated during folding and assembly but also that the various amino acid substitutions that inactivate the glycosylation consensus sequences do not themselves result in malfolding of the

chase

cr)

c

,e

N K)

x

TIZ

+1 Trypsin9 :-T--

+1-

92K--, -HAn 68K-

46K-

30K -

mutant proteins. To analyze the efficiency of transport of the mutant proteins to the cell surface, we did or did not treat intact monolayers of labeled, infected cells with trypsin during the last 15 min of chase as described above. Cell extracts were then prepared, and the HA molecules were immunoprecipitated and analyzed by SDS-PAGE. Figure 4 (upper panel) shows that in every case the great majority of the labeled proteins were cleaved by trypsin into HA1 and HA2 subunits, indicating that the mutants were transported to the cell surface during the 2-h chase as efficiently as the wild-type protein. Cleavage of the mutant polypeptides into their component HA1 and HA2 subunits also revealed the distinct increase in electrophoretic mobility resulting in each case from the loss of one of the oligosaccharide chains. Consistent with the position of each glycosylation consensus site in the polypeptide chain, mutants 8(-), 22(-), 38(-), 81(-), 165(-), and 285(-) show increased mobility of the HA1 subunit, while it is the HA2 subunit of mutant 483(-) that is

A2

30°C e

10

pulse 4h chase

y_

cc

i;

Trypsin]nI-1_+l 92K-

.

68K-

46K-

affected.

30K-

*

0*

42°C

C

Trypsinl

+

\j "'

+

aa

5 pulse

ri-,

4h chose

+-_~

cc-

a:

1-

c c": +

92K-

68K

41**" * * "

-~~~H a.

*~~~~~~~~~ *

46K-

30K2

FIG. 4. Analysis of the biosynthesis and intracellular transport of wild-type X31 HA and mutants lacking single glycosylation sites. CV-1 cells infected with recombinant viruses expressing wild-type or mutant X31 HA proteins were labeled with [35S]methionine as described in Materials and Methods for the times shown on the figure, at 37, 30 or 42°C. During the last 15 min of the chase, the intact monolayers were treated with medium (-) or medium containing 5 jLg of trypsin per ml (+) at the appropriate temperature before cell extracts were prepared and analyzed by immunoprecipitation and SDS-PAGE. Numbers on left show molecular mass in kilodaltons.

Although mutant HA molecules lacking a single glycosylation site showed no alteration in their competence for intracellular transport at 37°C, it was possible that absence of an oligosaccharide might result in a temperature-sensitive defect in transport. The wild-type and mutant HAs were therefore synthesized at either 30 or 42°C, and the intact cell monolayers were treated with trypsin during the last 15 min of the chase to detect HAO present on the cell membrane. The results shown in Fig. 4 (middle panel) demonstrate that at 30°C, all the mutant HAs move to the plasma membrane within a 4-h chase. The only difference that could be noted was that the HA1 subunit of mutant 165(-) is apparently more sensitive to protease compared with that of the other mutants or the wild-type protein. This suggests that when the protein is synthesized at 30°C, the loss of the oligosaccharide normally present at Asn-165 results in a structural alteration that exposes a protease-sensitive site(s) normally buried within the folded molecule. This is consistent with the observation that in the crystal structure of the X31 HA trimer, this oligosaccharide spans the interface between the globular domains at the top of the molecule, forming substantial contacts on the surfaces of the subunits (56). Increasing the temperature of synthesis to 42°C had a more dramatic effect on the transport and protease sensitivity of some of the mutant HA molecules (Fig. 4, lower panel). After 4 h of chase, the labeled HAO polypeptides of all the mutants except 285(-) became accessible to trypsin in the medium, demonstrating that they had reached the cell surface. Mutant 285(-) showed no evidence of an increase to a higher-molecular-mass, terminally glycosylated form of HAO, suggesting that its exit from the ER is blocked after synthesis at 42°C. The reason for the lack of transport of this mutant at 42°C may be related to the fact that it is the only

VOL. 66, 1992

;

A >

-

:

B

:

2h chase fDTT ..

. s

z

:ft,sw

R

--

0*4

4m

v~

GLYCOSYLATION OF INFLUENZA VIRUS HEMAGGLUTININ

-a a.

.~:

~

-.

92

7141

2h chase -DTT

co

*tft-A,rnl.

3 -- Nil c r~r> m

FIG. 5. Analysis of the biosynthesis, assembly, and intracellular transport of wild-type X31 HA and mutants having no or one glycosylation sites. (A) CV-1 cells infected with recombinant viruses expressing wild-type or mutant X31 HA proteins were labeled for 10 min with [35S]methionine and then incubated in medium containing an excess of nonradioactive methionine for 4 h at 37°C. The intact monolayers were then treated with medium (-) or medium containing 5 pLg of trypsin per ml (+) for 15 min at the appropriate temperature before cell extracts were prepared and analyzed by immunoprecipitation and SDS-PAGE. (B) Infected cells were labeled for 5 min with [35S]methionine and then incubated in medium containing an excess of nonradioactive methionine for 2 h at 37°C. Cell extracts were prepared and immunoprecipitated with a polyclonal anti-X31 HA antiserum. Immunoprecipitates were analyzed by SDS-PAGE under nonreducing conditions (-DTT) or after reduction (+D1T). Numbers on left show molecular mass in kilodaltons. Lane M.W., molecular size markers.

one in which glycine was substituted for serine in the glycosylation consensus sequence. The combination of the loss of the oligosaccharide and substitution of a residue that increases main-chain flexibility could result in an unstable molecule that cannot fold correctly at the elevated temperature. Several of the other mutants showed increased sensitivity to degradation by trypsin at 42°C. Both the HA1 and the HA2 subunits of mutants 81(-) and 165(-), which lack oligosaccharides normally located on the globular domain (Fig. 1), were almostly completely degraded by the protease. Mutants 8(-), 22(-), 38(-), and 483(-), all of which lack oligosaccharides normally present on the stem domain, appear to be less affected, although some degradation of both HA1 and HA2 subunits is apparent. Mutant HAs having lacking all seven or possessing a single oligosaccharide side chain are not transported to the plasma membrane and form aggregates in the ER. Analysis of the expression of the mutant (N) that lacks all seven of the carbohydrate attachment consensus sequences of the wildtype X31 HA molecule confirmed the result obtained by using tunicamycin to block glycosylation of HA and other newly synthesized polypeptides in the ER. Mutant N molecules labeled during a 10-min pulse at 37°C did not become accessible to trypsin in the medium after a 4-h chase (Fig. 5A). Synthesis at 30°C did not correct this transport defect (data not shown). To determine whether any one of the seven natural N-linked oligosaccharides would be sufficient to facilitate transport of the HA polypeptide to the plasma membrane, we analyzed a group of seven mutants [8(+), 22(+), 38(+), 81(+), 165(+), 285(+), and 483(+)] possessing only a single consensus site for glycosylation. These mutant proteins displayed shifts in electrophoretic mobility consistent with the absence of six oligosaccharide side chains when compared with the naked molecule and with wild-type X31 HA (Fig. 5A). By contrast to the group of mutants lacking only a single N-linked carbohydrate, mutant HAs possessing only a single glycosylation site do not become sensitive to cleavage into HA1 and HA2 subunits by trypsin in the medium after a 4-h chase at 37°C (Fig. 5A) or at 30°C (data not shown). The loss of signal in the HAO band of mutants 22(+), 38(+), 165(+), and 483(+) in immunoprecipitates from intact cell monolayers treated with trypsin before cell lysis is probably

due to incomplete removal of the protease before cell extracts were prepared since it was not consistently observed. To rule out the possibility that these molecules might be transported to the plasma membrane and then degraded by the protease, we determined the sensitivity of the mutant proteins to digestion with endo H. Infected cells were pulse-labeled for 10 min and chased for 4 h at 37°C, and then cell extracts were prepared and immunoprecipitated with

anti-HA serum. The immunoprecipitated proteins were then treated with endo H and analyzed by SDS-PAGE. The results of this analysis, summarized in Table 2, showed that none of the mutant proteins acquire resistance to endo H, suggesting that HA molecules bearing only a single oligosaccharide are retained within the ER and are not transported to the medial Golgi compartment where processing of mannose-rich, endo H-sensitive oligosaccharides to an endo H-resistant form occurs (26). The sharp, intense band migrating more slowly than the HA species is the ER resident chaperone, BiP, which coimmunoprecipitates with mutant HA molecules (12, 15, 25, 47). BiP interacts transiently with a variety of nascent wild-type exocytotic proteins (1, 2, 7, 28, 35), including HA (15), and more permanently with malfolded or unassembled proteins whose transport from the ER is blocked. The role played by BiP during the biosynthesis and folding of HA and other cell surface and secreted proteins is discussed in references 19 and 47. Several studies have indicated that correct folding and assembly of multimeric membrane glycoproteins occurs within the ER and is a prerequisite for transport through the exocytic pathway (3, 4, 10, 15, 25, 27) and that misfolded proteins form disulfide-bonded aggregates in that compartment (25, 30, 47). To test the aggregation state of the ER-blocked mutants, immunoprecipitates from extracts of infected cells labeled for 2 h were analyzed by SDS-PAGE under reducing and nonreducing conditions. As seen in Fig. SA, the mutant proteins migrate under reducing conditions as single species that are well separated from the coimmunoprecipitated BiP molecules. By contrast, the mutant HA molecules migrate under nonreducing conditions as a heterogeneous smear near the top of the gel, suggesting that they form large, disulfide-linked aggregates (Fig. 5B). The pattern of trimers, dimers, and monomers, diagnostic of the forma-

7142

GALLAGHER ET AL.

J. VIROL.

TABLE 2. Endo H sensitivity, erythrocyte binding, and cell fusion characteristics of wild-type and mutant HA proteinsa

m rr

I

HA

Endo Hb

Erythrocyte binding

Cell

M%)C >90

fusion'

+++ +++ +++ +++ +++ +++ +++

X31 HA

R

8(-) 22(-) 38(-) 81(-) 165(-) 285(-) 483(-)

R R R R R R R

>90 >90 >90 >90 >90

N 8(-)

S

0 0 0 0 0 0 0 0

-

-

>90 >90

C) 0

sE o

ur o

ur o

0 -0

0

I

I

FC-- +-i- +1-

I

+1-

c) =0< I I 0 >x

+1-

+!

+++ BiP

] HA,

-

R ] HA-

0 HA2

22(+) 38(+) 81(+) 165(+) 285(+) 483(+)

S S S S

S S S

CHO(+3) CHO(+4)

S

0 0

CHO(+5)A CHO(+5)B CHO(+5)C

R R R

80 80 80

S

-

+++ +++ +++

a The synthesis of HA glycoproteins at 37 or 30'C in CV-1 cells infected with recombinant viruses and the assays to determine sensitivity of oligosaccharide side chains to endo H, erythrocyte binding, and polykaryon formation (cell fusion) were done as described in Materials and Methods. b R, acquisition of resistance to endo H digestion observed for some oligosaccharide side chains after 4 h of chase at 37°C. S, all oligosaccharide side chains were sensitive to digestion by endo H. c Determined at both 30 and 37°C.

tion of native trimers by the wild-type protein and the single oligosaccharide deletion mutants (Fig. 3), is absent, indicating that mutants having one or fewer oligosaccharide side chains are unable to assemble to form native structures. As observed previously (25, 47), the electrophoretic mobility of the coimmunoprecipitated BiP molecules is not altered under nonreducing conditions (Fig. SB), indicative of the noncovalent nature of the interaction between the mutant HA and BiP proteins. Mutant HA molecules having only five oligosaccharide side chains are transported to the plasma membrane. To determine whether various combinations of subsets of the seven natural oligosaccharide side chains would be sufficient for transport of X31 HA to the plasma membrane, we examined a third group of mutant HAs. This group, composed of five mutants possessing three [CHO(+3)], four [CHO(+4)], or three combinations of five [CHO(+5)A, CHO(+5)B, CHO(+5)C] glycosylation sites (Fig. 1), was analyzed for transport to the plasma membrane. Figure 6 shows that mutants having any of the three combinations of five glycosylation sites are transported to the plasma membrane as evidenced by the appearance of HA1 and HA2 subunits when intact monolayers were treated with trypsin before lysis and immunoprecipitation. The efficiency with which these mutant HAs [particularly CHO(+5)B and CHO(+5)C] move to the plasma membrane appears to be reduced compared with that of wild-type X31 HA, since a smaller proportion of the labeled mutant proteins were accessible to the protease at the end of the 4-h chase at 37°C. Identical results were

FIG. 6. Analysis of the biosynthesis, assembly, and intracellular transport of wild-type X31 HA and mutants having various combinations of glycosylation sites. CV-1 cells infected with recombinant viruses expressing wild-type or mutant X31 HA proteins were labeled for 10 min with [35S]methionine and then incubated in medium containing an excess of nonradioactive methionine for 4 h at 37°C. The intact monolayers were then treated with medium (-) or medium containing 5 ,ug of trypsin per ml (+) for 15 min at the appropriate temperature before cell extracts were prepared and analyzed by immunoprecipitation and SDS-PAGE. Lane M.W., molecular weight markers.

obtained when the mutants were synthesized at 30°C (data not shown). Significant amounts of BiP were observed to coprecipitate with these mutants, but not with the wild-type

HA, suggesting that the less efficient folding and assembly of the mutant proteins results in accumulation of some malfolded, BiP-associated molecules in the ER (Fig. 6). However, electrophoresis of these mutants under nonreducing conditions demonstrated a significant amount of trimeric molecules (data not shown). The CHO(+3) and CHO(+4) mutants never became accessible to trypsin in the medium, indicating that they were not transported to the cell surface. Again, significant amounts of BiP coprecipitated with the mutant proteins. In an experiment identical to that shown in Fig. SB, the CHO(+3) and CHO(+4) proteins migrated during electrophoresis under nonreducing conditions only as a heterogeneous smear near the top of the gel (data not shown), suggesting that they form large, disulfide-linked aggregates in the ER. Consistent with the results of the transport studies, the majority of labeled mutant HA molecules having five glycosylation sites acquired resistance to digestion with endo H during a 4-h chase, while the two mutant proteins having either three or four glycosylation sites did not (Table 2). Identical results were obtained when these mutants were synthesized at 30°C (data not shown), indicating that synthesis at the lower temperature could not correct the transport defect. Glycosylation mutants that are transported to the cell surface retain the biological activities of the wild-type HA molecule. The presence of functionally active HA on the cell surface can be detected by using two assays that reflect the receptor-binding and membrane fusion properties of the molecule (17). The sialic acid-binding activity of HA can be monitored by hemagglutination of erythrocytes to infected cell monolayers (16). CV-1 cells expressing HA on their plasma membranes can be induced to fuse with each other

VOL. 66, 1992

GLYCOSYLATION OF INFLUENZA VIRUS HEMAGGLUTININ

after brief exposure to low pH (14, 52), resulting in the formation of large syncitial cells. Table 2 summarizes the results obtained when cells infected for 40 h at 37 or 30°C with recombinant viruses encoding wild-type or mutant X31 HAs were assayed for erythrocyte-binding fusion activity. Cells expressing mutant HAs having six or five glycosylation sites agglutinated erythrocytes and formed large syncytia after brief exposure to low pH, indicating that these mutants are displayed on the cell surface in a biologically active conformation. By contrast, cells expressing mutant HAs having less than five N-linked carbohydrate side chains were not observed to bind erythrocytes or undergo low pHinduced fusion. DISCUSSION Our analysis of the role of the seven N-linked oligosaccharide side chains on X31 HA has demonstrated that no single side chain is essential for the folding, assembly, intracellular transport, or biological activity of the molecule under normal physiological conditions. However, biosynthetic studies at lower and higher temperatures revealed some differences in the efficiency of folding and/or stability of individual mutants lacking one oligosaccharide. Interestingly, the two mutants missing the side chains that form the most intimate contacts with the folded polypeptide chain displayed the most significant defects. These side chains, which are attached to asparagine residues 165 and 285, are the only ones that are (i) clearly visible in electron density maps of X31 HA crystals (56) and (ii) protected by the folding of the polypeptide chain from being processed into complex oligosaccharides. That a double mutant lacking both these side chains was folded and transported to the cell surface, albeit less efficiently than the wild-type protein, indicates that despite their apparent inclusion as an integral part of the wild-type HA structure, both sites can be dispensed with without serious consequences. A single oligosaccharide present at any one of the seven attachment sites was not alone sufficient to promote the folding of the X31 HA molecule. Each mutant protein containing only one side chain, like the mutant lacking all carbohydrate attachment sites and the nonglycosylated wildtype protein synthesized in the presence of tunicamycin (25), accumulated in the ER as large disulfide-bonded aggregates that were noncovalently associated with the ER chaperone BiP. Mutants containing just the four side chains located on the stem domain, or the three side chains located on the globular domain and hinge region, also accumulated as aggregates in the ER, indicating that both portions of the molecule require at least some carbohydrate for folding and stabilization. However, three mutants containing different combinations of five oligosaccharide chains were folded and transported to the cell surface, albeit less efficiently than the wild-type protein. It is of interest that a previous analysis of nonglycosylated X31 HA synthesized in the presence of tunicamycin with antibodies that distinguish native and denatured epitopes indicated that the stem region is misfolded, while the globular domain appeared to have a native structure (25). These observations suggest that the absence of oligosaccharides on the stem domain might be the major cause of misfolding, and our analysis of mutant CHO(+5)C showed that the presence of the subset of five oligosaccharides on the stem domain and hinge region was indeed sufficient to facilitate folding of the molecule. Nevertheless, although it was not feasible to test every possible combination of different numbers of side chains, it appears that the

7143

presence of a majority (five or more) of the side chains, but no particular combination of them, is required for the protein to fold into a transport-competent structure. Our results support the currently accepted view that carbohydrate side chains are important for maintenance of the structure and stability of glycoproteins (36). Several other studies have shown that nonglycosylated or underglycosylated membrane and secretory proteins are defective in folding and intracellular transport and/or vulnerable to degradation (20, 22, 29-32, 34, 40, 49), although some others appear relatively unaffected (20, 23, 36, 55). As is particularly apparent from studies of pairs of homologous proteins, the degree of need for glycosylation is an intrinsic property dictated by the primary amino sequence of each individual glycoprotein. Thus, Japan HA, like the G protein of the Orsay strain of vesicular stomatitis virus (21), is relatively insensitive to the effects of tunicamycin, while their homologs X31 HA and the G protein from the San Juan strain of vesicular stomatitis virus (21, 29, 38) have quite stringent requirements for glycosylation. In the San Juan G protein, the necessity for glycosylation at least one of the two natural attachment sites is determined by the presence of a tyrosine at residue 172, whose replacement by an aspartic acid (the equivalent residue in Orsay G) by site-directed mutagenesis results in a polypeptide having the less stringent glycosylation requirement of the Orsay protein (38). This result supports the idea that the function of individual oligosaccharide side chains is to incrementally increase the stability of the folded form of the molecule, either via direct interactions with the polypeptide or by shielding potentially interactive surfaces. In some proteins, for example, the HA-neuraminidase glycoprotein of SV5, individual side chains can be identified that play major roles in promoting correct folding or in preventing aggregation (34). By contrast, in X31 HA, no single oligosaccharide is sufficient or necessary alone to stabilize the protein structure. Rather, the accumulated effect of five or more of the seven side chains is required to facilitate the folding and transport of the molecule. If five or six oligosaccharides are sufficient for X31 HA to achieve a conformation that satisfies the structural requirement for transport to the cell membrane, then why has this protein evolved to have seven glycosylation sites? Our analysis of HA in a tissue culture system that does not include the other viral components clearly does not reproduce all the challenges faced by the influenza virion in an infected animal. It has already been established that one function of individual oligosaccharide side chains is to shield antigenic regions of the HA polypeptide from detection by the host's immune system (48). In addition, the positioning of the oligosaccharides may modulate the efficiency of cleavage activation of HA and thus the virulence of the virus infection (5). It is also possible that the additional carbohydrate may promote or prevent intermolecular interactions that facilitate or hinder efficient incorporation of HA into virions. Testing of these possibilities will not be possible until it is routinely feasible to reintroduce altered HA mole-

cules into influenza virus particles. ACKNOWLEDGMENTS We thank Carolyn Doyle, Mike Roth, Karen McCammon, and Tom Bittick for helpful discussions and assistance. This work was funded by grants from the American Cancer Society and the National Institutes of Health to M.-J.H.G. and from the Welch Foundation to J.F.S. P.J.G. was supported by a research fellowship awarded by the Damon Runyon-Walter Winchell Cancer Foundation.

7144

1. 2.

3.

4.

5.

6. 7.

8.

9.

10. 11. 12.

13.

14.

15. 16. 17.

18. 19. 20. 21.

J. VIROL.

GALLAGHER ET AL.

REFERENCES Blount, P., and J. P. Merlie. 1991. BiP associates with newly synthesized subunits of the mouse muscle nicotinic receptor. J. Cell Biol. 113:1125-1132. Bole, D. G., L. M. Hendershot, and J. F. Kearney. 1986. Posttranslational association of immunoglobulin heavy chain binding protein with nascent heavy chains in nonsecreting and secreting hybridomas. J. Cell Biol. 102:1558-1566. Copeland, C. S., R. W. Doms, E. M. Bolzau, R. G. Webster, and A. Helenius. 1986. Assembly of influenza hemagglutinin trimers and its role in intracellular transport. J. Cell Biol. 103:11791191. Copeland, C. S., K.-P. Zimmer, K. R. Wagner, G. A. Healey, I. Mellman, and A. Helenius. 1988. Folding, trimerization and transport are sequential events in the biogenesis of influenza virus hemagglutinin. Cell 53:197-209. Deshpande, K. L., V. A. Freid, M. Ando, and R. G. Webster. 1987. Glycosylation affects cleavage of an HSN2 influenza virus hemagglutinin and regulates virulence. Proc. Natl. Acad. Sci. USA 84:36-40. Doms, R. W., and A. Helenius. 1986. Quaternary structure of influenza virus hemagglutinin after acid treatment. J. Virol. 60:833-839. Dorner, A. J., D. G. Bole, and R. J. Kaufmnan. 1987. The relationship of N-linked glycosylation and heavy chain-binding protein association with the secretion of glycoproteins. J. Cell Biol. 105:2665-2674. Doyle, C., M. G. Roth, J. Sambrook, and M. J. Gething. 1985. Mutations in the cytoplasmic domain of the influenza virus haemagglutinin affect different stages of intracellular transport. J. Cell Biol. 100:704-714. Doyle, C., J. Sambrook, and M. J. Gething. 1986. Analysis of progressive deletions of the transmembrane and cytoplasmic domains of influenza hemagglutinin. J. Cell Biol. 103:1193-1204. Einfeld, D., and E. Hunter. 1988. Oligomeric structure of a prototype retrovirus glycoprotein. Proc. Natl. Acad. Sci. USA 85:8688-8692. Elbein, A. D. 1987. Inhibitors of the biosynthesis and processing of n-linked oligosaccharide chains. Annu. Rev. Biochem. 56: 497-534. Gallagher, P., J. Henneberry, I. Wilson, J. Sambrook, and M. J. Gething. 1988. Addition of carbohydrate side chains at novel sites on influenza virus hemagglutinin can modulate the folding, transport and activity of the molecule. J. Cell Biol. 107:20592073. Gething, M. J., J. Bye, J. J. Skehel, and M. Waterfield. 1980. Cloning and DNA sequence of double-stranded copies of haemagglutinin genes from H2 and H3 strains elucidates antigenic shift and drift in human influenza virus. Nature (London) 287:301-306. Gething, M. J., R. W. Doms, D. York, and J. M. White. 1986. Studies on the mechanism of membrane fusion: site-specific mutagenesis of the hemagglutinin of influenza virus. J. Cell Biol. 102:11-23. Gething, M. J., K. McCammon, and J. Sambrook. 1986. Expression of wild-type and mutant forms of influenza hemagglutinin: the role of folding in intracellular transport. Cell 46:939-950. Gething, M. J., and J. Sambrook. 1981. Cell-surface expression of influenza haemagglutinin from a cloned DNA copy of the RNA gene. Nature (London) 293:620-625. Gething, M. J., and J. SambrooL 1983. Expression of cloned influenza virus genes, p. 169-191. In P. Palese and D. W. Kingsbury (ed.), Genetics of influenza viruses. Springer-Verlag KG, Vienna. Gething, M. J., and J. Sambrook. 1989. Protein folding and intracellular transport: studies on influenza virus haemagglutinin. Biochem. Soc. Symp. 55:155-166. Gething, M. J., and J. F. Sambrook. 1992. Protein folding in the cell. Nature (London) 355:33-45. Gibson, R., S. Kornfeld, and S. Schlesinger. 1980. A role for oligosaccharides in glycoprotein biosynthesis. Trends Biochem. Sci. 290-293. Gibson, R., S. Schlesinger, and S. Kornfeld. 1979. The nongly-

cosylated glycoprotein of vesicular stomatitis virus is temperature-sensitive and undergoes intracellular aggregation at elevated temperatures. J. Biol. Chem. 254:3600-3607. 22. Guan, J.-L., H. Cao, and J. K. Rose. 1988. Cell-surface expression of a membrane-anchored form of the human chorionic gonadotropin a subunit. J. Biol. Chem. 263:5306-5313. 23. Hannink, M., and D. J. Donoghue. 1986. Cell surface expression of membrane-anchored v-sis gene products: glycosylation is not required for cell surface transport. J. Cell Biol. 103:2311-2322. 24. Hirschberg, C. B., and M. D. Snider. 1987. Topography of glycosylation in the rough endoplasmic reticulum and Golgi apparatus. Annu. Rev. Biochem. 56:63-87. 25. Hurtley, S. M., D. G. Bole, H. Hoover-Litty, A. Helenius, and

26. 27. 28.

29.

30.

31.

C. S. Copeland. 1989. Interactions of misfolded influenza virus hemagglutinin with binding protein (BiP). J. Cell Biol. 108:21172126. Kornfeld, R., and S. Kornfeld. 1985. Assembly of asparaginelinked oligosaccharides. Annu. Rev. Biochem. 54:631-664. Kreis, T. E., and H. F. Lodish. 1986. Oligomerization is essential for transport of vesicular stomatitis viral glycoprotein to the cell surface. Cell 46:929-937. Machamer, C. E., R. W. Doms, D. G. Bole, A. Helenius, and J. K. Rose. 1990. Heavy chain binding protein recognizes incompletely disulfide-bonded forms of vesicular stomatitis virus G protein. J. Biol. Chem. 265:6879-6883. Machamer, C. E., R. Z. Florkiewicz, and J. K. Rose. 1985. A single N-linked oligosaccharide at either of the two normal sites is sufficient for transport of vesicular stomatitis virus G protein to the cell surface. Mol. Cell. Biol. 5:3074-3083. Machamer, C. E., and J. K. Rose. 1988. Vesicular stomatitis virus G proteins with altered glycosylation sites display temperature-sensitive intracellular transport and are subject to aberrant intermolecular disulfide bonding. J. Biol. Chem. 263:59555960. Matzuk, M. M., and I. Boime. 1988. The role of the asparaginelinked oligosaccharides of the a subunit in the secretion and

assembly of human chorionic gonadotropin. J. Cell Biol. 106: 1049-1059. 32. Miyazaki, J.-I., E. Appella, H. Zhao, J. Forman, and K. Ozato. 1986. Expression and function of a nonglycosylated major histocompatibility class I antigen. J. Exp. Med. 163:856-871. 33. Nakamura, K., and R. W. Compans. 1979. Host cell- and virus

strain-dependent differences in oligosaccharides of hemagglutinin glycoproteins of influenza A viruses. Virology 95:8-23. 34. Ng, D. T. W., S. W. Hiebert, and R. A. Lamb. 1990. Different roles of individual N-linked oligosaccharide chains in folding, assembly, and transport of the simian virus 5 hemagglutininneuraminidase. Mol. Cell. Biol. 10:1989-2001. 35. Ng, D. T. W., R. E. Randall, and R. A. Lamb. 1989. Intracellular maturation and transport of the SV5 type II glycoprotein hemagglutinin-neuraminidase: specific and transient association with GRP78-BiP in the endoplasmic reticulum and extensive internalization from the cell surface. J. Cell Biol. 109:3273-3289. 36. Olden, K., J. B. Parent, and S. L. White. 1982. Carbohydrate moieties of glycoproteins. A re-evaluation of their function. Biochem. Biophys. Acta 650:209-232. 37. Owen, M. J., A.-M. Kissonerghis, and H. F. Lodish. 1980. Biosynthesis of HLA-A and HLA-B antigens in vivo. J. Biol. Chem. 255:9678-9684. 38. Pitta, A. M., J. K. Rose, and C. E. Machamer. 1989. A single amino acid substitution eliminates the stringent carbohydrate requirement for intracellular transport of a viral glycoprotein. J. Virol. 63:3801-3809. 39. Rademacher, T. W., R. B. Parekh, and R. A. DweL 1992. Glycobiology. Annu. Rev. Biochem. 57:785-838. 40. Rogers, M., A. Taraboulos, M. Scott, D. Groth, and S. B. Prusiner. 1990. Intracellular accumulation of the cellular prion protein after mutagenesis of its Asn-linked glycosylation sites. Glycobiology 1:101-109. 41. Roth, M. G., M. J. Gething, and J. Sambrook. 1989. Membrane insertion and intracellular transport of influenza virus glycoproteins, p. 219-267. In R. M. Krug (ed.), The viruses. Plenum Press, New York.

VOL. 66, 1992

GLYCOSYLATION OF INFLUENZA VIRUS HEMAGGLUTININ

42. Rott, R., M. Orlich, H.-D. Kienk, M. L. Wang, J. J. Skehel, and D. C. Wiley. 1984. Studies on the adaptation of influenza viruses to MDCK cells. EMBO J. 3:3329-3332. 43. Sambrook, J., E. F. Fritsch, and T. Maniatis. 1989. Molecular cloning: a laboratory manual, 2nd ed. Cold Spring Harbor Laboratory, Cold Spring Harbor, N.Y. 44. Sanger, F., S. Nicklen, and A. R. Coulson. 1977. DNA sequencing with chain-terminating inhibitors. Proc. Natl. Acad. Sci. USA 74:5463-5467. 45. Santos-Aguado, J., P. A. Biro, U. Fuhrmann, J. L. Strominger, and J. A. Barbosa. 1987. Amino acid sequences in the arl domain and not glycosylation are important in HLA-A2/32-microglobulin association and cell surface expression. Mol. Cell. Biol. 7:982-990. 46. Schulz, G. E., and R, H. Schirmer. 1979. Principles of protein structure. Springer-Verlag, New York. 47. Segal, M. S., J. M. Bye, J. F. Sambrook, and M. J. Gething. 1992. Disulfide bond formation during the folding of influenza virus hemagglutinin. J. Cell Biol. 118:227-244. 48. Skehel, J. J., D. J. Stevens, R. S. Daniels, A. R. Douglas, M. Knossow, I. A. Wilson, and D. C. Wiley. 1984. A carbohydrate side chain on hemagglutinins of Hong Kong influenza viruses inhibits recognition by a monoclonal antibody. Proc. Natl. Acad. Sci. USA 81:1779-1783. 49. Taylor, A. K., and R. Wall. 1988. Selective removal of alpha heavy-chain glycosylation sites causes immunoglobulin A degradation and reduced secretion. Mol. Cell. Biol. 8:4197-4203. 50. Ward, C. 1981. Structure of the influenza virus hemagglutinin. Curr. Top. Microbiol. Immunol. 94/95:2-66. 51. Waterfield, M. D., M. J. Gething, G. Scrace, and J. J. Skehel.

52. 53.

54.

55. 56.

57.

58.

7145

1980. The carbohydrate side chains and disulphide bonds of the haemagglutinin of the influenza virus A/Japan 305/57 (H2N1), p. 11-20. In G. Laver and G. Air (ed.), Structure and variation in influenza virus. Elsevier/North Holland, Australia. White, J. M., A. Helenius, and M. J. Gething. 1982. Haemagglutinin of influenza virus expressed from a cloned gene promotes membrane fusion. Nature (London) 300:658-659. Wiley, D. C., and J. J. Skehel. 1987. The structure and function of the hemagglutinin membrane glycoprotein of influenza virus. Annu. Rev. Biochem. 56:365-394. Williams, M. A., and R. A. Lamb. 1986. Determination of the orientation of an integral membrane protein and sites of glycosylation by oligonucleotide-directed mutagenesis: influenza B virus NB glycoprotein lacks a cleavable signal sequence and has an extracellular NH2-terminal region. Mol. Cell. Biol. 6:43174328. Williams, M. A., and R. A. Lamb. 1988. Polylactosaminoglycan modification of a small integral membrane glycoprotein, influenza B virus NB. Mol. Cell. Biol. 8:1186-1196. Wilson, I. A., J. J. Skehel, and D. C. Wiley. 1981. Structure of the haemagglutinin membrane glycoprotein of influenza virus at 3 A resolution. Nature (London) 289:366-373. Yewdell, J. W., A. Yellen, and T. Bachi. 1988. Monoclonal antibodies localize events in the folding, assembly, and intracellular transport of the influenza virus hemagglutinin glycoprotein. Cell 52:843-852. Zoller, M. J., and M. Smith. 1985. Oligonucleotide-directed mutagenesis: a simple method using two oligonucleotide primers and a single-stranded DNA template. DNA 3:479-488.