Effect of Hemagglutinin Glycosylation on Influenza
http://www.100md.com
病菌学杂志 2005年第19期
Department of Internal Medicine, School of Medicine, University of Virginia, Charlottesville, Virginia
Department of Pathology, School of Medicine, University of Virginia, Charlottesville, Virginia
ABSTRACT
Inhibition of neuraminidase (NA) activity prevents release of progeny virions from influenza-infected cells and removal of neuraminic (sialic) acid moieties from glycans attached to hemagglutinin (HA). Neuraminic acid moieties situated near the HA receptor-binding site can reduce the efficiency of virus binding and decrease viral dependence on NA activity for replication. With the use of reverse genetics technique, we investigated the effect of glycans attached at Asn 94a, 129, and 163 on the virus susceptibility to NA inhibitors in MDCK cells and demonstrated that the glycan attached at Asn 163 plays a dominant role in compensation for the loss of NA activity.
INTRODUCTION
The two surface glycoproteins, hemagglutinin (HA) and neuraminidase (NA), of influenza A virus interact with cellular receptors containing terminal neuraminic acid (NeuAc) moieties. HA initiates infection by binding to cellular receptors, whereas NA destroys the receptors by cleaving off NeuAc moieties (49). In addition, HA binds to complementary, as-yet-unidentified, cellular receptors that lack NeuAc moieties (5, 46). The functional balance between HA and NA in influenza virus infections has been intensively studied (4, 10, 22, 23, 33, 49, 50). The HA precursor (HA0) is proteolytically cleaved into two subunits, HA1 and HA2; the HA1 subunit carries the NeuAc-binding site, and the HA2 subunit is responsible for fusion of viral and cellular membranes (51). NA is not essential for influenza A virus assembly or budding (25), but its enzymatic activity facilitates the progeny virions' release from infected cells (6). The anti-influenza drugs oseltamivir and zanamivir inhibit NA activity by targeting the enzyme active site formed by highly conserved residues (6, 24). Substitutions at those residues confer virus resistance to NA inhibitors (28). However, drug resistance can also be conferred via an NA-independent mechanism. Amino acid substitutions in HA can lessen viral dependence on NA activity for release from infected cells and thus decrease susceptibility to NA inhibitors by reducing efficiency of virus binding to cellular receptors (11, 29, 41, 49).
Propagation of virus in the presence of an NA inhibitor should result in retention of NeuAc moieties by the complex glycans attached to HA (3, 39). Of note, negatively charged NeuAc moieties situated near the HA receptor-binding site have a potential to impair the virus binding to cellular receptors (38). Therefore, viruses containing complex glycans at the receptor-binding site would exhibit lower susceptibility to NA inhibitors than those lacking such glycans. To investigate this possibility, we utilized the influenza A viruses that belong to the H1N1 antigenic subtype as a model. Viruses of this subtype produced the devastating "Spanish flu" pandemic in 1918 (42), and antigenic drift variants of this virus remained in circulation in the human population until mid-1950s. In 1977, the virus reemerged in the human population, causing the "Russian flu" epidemic; since that time, antigenic drift variants have been in circulation in the human population. There is a considerable variance in the glycosylation patterns near the HA receptor-binding site among the human influenza A/H1N1 viruses, due to antigenic drift and host adaptation (20). For example, the so-called early laboratory-passaged virus A/WSN/33 (WSN) contains a single glycosylation site at Asn129, whereas the contemporary virus, A/Charlottesville/31/95 (CH/95), contains two additional glycosylation sites at Asn 94a and Asn 163) (Fig. 1) (numbering according to reference 52). In our previous studies, we demonstrated that both viruses were equally susceptible to NA inhibitors by the enzyme inhibition assay, whereas in cell culture, the CH/95 virus was drug resistant and the WSN virus was drug susceptible (12). Moreover, reassortant virus that contained the HA gene from the CH/95 virus in the WSN virus genetic background also exhibited drug resistance in cell culture (12). In the present study, we investigated the contribution of HA glycans attached at Asn 94a, 129, and 163 to viral dependence on NA activity. Our findings provide further insights into molecular mechanisms underlying the human virus resistance to NA inhibitors in cell culture.
MATERIALS AND METHODS
Compounds. Inhibitors of influenza virus neuraminidase activity zanamivir (GlaxoWellcome Research and Development, Stevenage, United Kingdom) and oseltamivir (Roche Laboratories, Inc., Nutley, NJ) were provided by their respective manufacturers. Inhibitors of N-glycosylation 1-deoxymannojirimycin (dMj) and swainsonine were purchased from Sigma, St. Louis, MO.
Viruses and cells. The A/Charlottesville/31/95 (H1N1) (CH/95) virus was from the repository of the Respiratory Disease Study Unit at the University of Virginia. The recombinant A/WSN/33 (H1N1) virus and its reassortants containing the HA gene from the CH/95 virus were generated by the reverse genetics approach (see below). Virus stocks prepared in Madin-Darby canine kidney (MDCK) cells were stored at –80°C. MDCK cells (a kind gift of Alan Hay, Mill Hill, United Kingdom) were utilized in all experiments, including determinations of virus titer and assessment of drug susceptibility. MDCK cells were grown in 24-well plates in Eagle's minimum essential medium (EMEM) (Cambrex, Wallorsville, MD) supplemented with 10% fetal calf serum (HyClone, Logan, UT) and antibiotics. Before infection, cell monolayers were rinsed twice with Dulbecco's phosphate-buffered saline (DPBS) to remove serum. Virus was inoculated onto confluent cellular monolayers and after 1 h of adsorption at room temperature, the cells were washed with DPBS to remove unbound virus. EMEM supplemented with N-tosyl-L-phenylalanine chloromethyl ketone (TPCK)-treated trypsin (Worthington, Lakewood, NJ) to a final concentration of 1 μg/ml was added to cells. The cells were incubated at 37°C for 48 h, at which point cell culture supernatants were harvested and used to assess virus titers in MDCK cells by a standard assay (16).
Plasmids construction. With the use of PCR-directed mutagenesis and high-fidelity KOD DNA polymerase (Novagen, Madison, WI), plasmid pHW-HA31 encoding the HA gene of CH/95 virus (12) was modified to abolish a glycosylation motif (Asn-X-Thr/Ser) at residues 94a, 129, and 163. To remove a glycosylation site at residue 94a, the sequence AAT-GGA-ACA (Asn-Gly-Thr) was changed to CAG-GGA-ACA (Gln-Gly-Thr). To remove sites at position 129, the sequence AAC-CAC-ACC (Asn-His-Thr) was changed to CAG-CAC-ACC (Gln-His-Thr), and at residue 163, AAT-CTG-AGC (Asn-Leu-Ser) was changed to CAG-CTG-AGC (Gln-Leu-Ser) (Table 1). In addition, plasmid pHW184 (18) encoding the HA gene (WSN) was modified to create a glycan attachment site at residue 94a and/or 163. The sequence AAT-GGA-GCA (94aAsn-Gly-Ala) was changed to AAT-GGA-ACA (94aAsn-Gly-Thr) and/or the sequence AAG-CTG-ACC (163Lys-Leu-Thr) was changed to AAT-CTG-AAC (163Asn-Leu-Thr). Lack of unwanted mutations was confirmed by sequence analysis of the plasmid DNAs.
Sequence analysis. Extraction of viral RNA from MDCK cell supernatants was performed with the use of RNAeasy kit according to the manufacturer's manual (QIAGEN, Clarita, CA). The primer U9 5'-AGCAGAAGC-3' was utilized to generate cDNA with the use of MMLV reverse transcriptase (Promega, Madison, WI). The HA gene was then amplified with a use of a standard PCR and the PCR product was then purified with the use of the QIAquick PCR purification kit according to the manufacturer's instructions (QIAGEN, Clarita, CA). Purified PCR product was subjected to sequence analysis with the use of ABI 373 DNA sequencer (Applied Biosystems, Foster City, CA) at the Center of Biotechnology at the University of Virginia. Sequencher 4.0 software (Gene Codes Corporation, Ann Arbor, MI) was used for analysis and translation of nucleotide sequence data.
Procedure for generation of HA glycosylation site variants. To generate recombinant viruses, the eight-plasmid reverse genetics system that allows synthesis of both templates vRNA (negative sense) and mRNA (positive sense) from the same plasmid was utilized (19). Coculture of 293T human embryonic kidney cells and MDCK cells was maintained in six-well plates in EMEM supplemented with 10% fetal calf serum. Cells were washed with DPBS, and then OptiMEM medium (Invitrogen, Carlsbad, CA) supplemented with antibiotics and antimycotics was added. Cells were then transfected with a mixture of eight plasmid DNAs (total of 1 μg/well) by using TransIt transfection reagent (Panvera, Madison, WI) as previously described (18). When specified, plasmid encoding a modified HA gene (either WSN or CH/95) was used instead of plasmid encoding the wild-type HA (WSN). At 24 h posttransfection, cell culture supernatant was replaced with EMEM supplemented with 0.3% bovine serum albumin (BSA; Gibco, Carlsbad, CA) and TPCK-treated trypsin at a final concentration of 1 μg/ml. Supernatants containing the recombinant viruses were collected 48 h later and used to infect MDCK cell monolayers by a standard procedure. Lack of unwanted mutations in HA genes of recombinant viruses was confirmed by the sequence analysis.
Radioactive labeling of viral proteins and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis. MDCK cell monolayers were infected with virus at a multiplicity of infection (MOI) of 5 50% tissue culture infective doses (TCID50) per cell. At 4 h postinfection, the cells were metabolically radiolabeled with [35S]methionine-cysteine (100 μCi per 106 cells) (Perkin-Elmer, Boston, MA). A 15-min pulse was followed by a 45-min chase, and then the supernatants were discarded and the cells were lysed in RIPA buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% IGEPAL CA-630, 0.5% sodium deoxycholate, 0.1% SDS, and 2 mM EDTA). The precursor HA0 protein was immunoprecipitated with anti-HA monoclonal antibodies (MAb) IVC102. To remove N-glycosylated moieties, an aliquot of each immunoprecipitate was pretreated with of peptide:N-glycosidase F (PNGase F; New England Biolabs, Beverly, MA) according to the manufacturer's recommendations for 6 h at 37°C.
To analyze the effect of inhibitors of N-glycosylation on the oligosaccharide composition of the HA molecules incorporated into the virus particles, MDCK cells were infected at an MOI of 5; after 1 h of virus adsorption, EMEM containing either dMj (Sigma, St. Louis, MO) or swainsonine (Sigma, St. Louis, MO) at a final concentration of 75 μM (15 μg/ml) and 60 μM (10 μg/ml), respectively, was added. At 4 h postinfection, the supernatants were removed and the cells were metabolically radiolabeled with [35S]methionine-cysteine (100 μCi per 106) in the presence of either dMj or swainsonine. At 18 h postinfection, the cell culture supernatants were harvested, and the cellular debris was removed by centrifugation at 12,000 rpm in an Eppendorf centrifuge for 15 min and then subjected to ultracentrifugation through a 25% sucrose cushion at 27,000 rpm (Beckman centrifuge, rotor SW28) for 90 min. The pellet containing the virus particles was resuspended in 50 μl of a loading gel buffer containing SDS. The denatured proteins were then incubated with either endo H glycosidase (500 U; New England Biolabs, Beverly, MA) or PNGase F for 12 h. At the end of the digestions, the samples and the control incubations without enzyme were examined by 8% SDS-PAGE, followed by fluorography.
Rescue procedure for recombinant NA activity-lacking virus variants. Generation of viruses lacking NA activity was done by transfection of 293T cells cocultured with MDCK cells, essentially as described above. The plasmid encoding the wild-type NA (WSN) was replaced with the plasmid (pdelNA/eGFP) that encodes the enhanced green fluorescence protein (eGFP) instead of the NA active site (32, 44). Virus spread was monitored in live cells with the use of Zeiss Axiovert 135 TV inverted microscope equipped with a cooled charge-coupled device camera (Hamamatsu Orca, Japan) and with excitation (488 nm) and emission (530 nm) filters (ChromaTechnology Corp., Brattleboro, VT). At 72 h posttransfection (48 h after addition of trypsin to the medium), the supernatants were harvested and used to inoculate fresh monolayers of MDCK cells. Rescue was considered successful if cytopathic effect (CPE) was observed on day 3 postinfection and HA activity was detected in MDCK cell supernatants.
Rescue of NA-lacking virus variants in the presence of inhibitors of N-glycosylation. Three hours after transfection, cell culture supernatants were removed and cells were washed with DPBS. OptiMEM containing either dMj (Sigma, St. Louis, MO) at a final concentration of 75 μM (15 μg/ml) or swainsonine (Sigma, St. Louis, MO) at a final concentration of 60 μM (10 μg/ml) was added. These concentrations were reported to produce desirable effects on composition of oligosaccharide chains of glycoproteins synthesized in various cell cultures including MDCK cells (8, 9, 17, 40). At 24 h posttransfection, supernatants were replaced with fresh EMEM containing TPCK-trypsin and the corresponding inhibitor of N-glycosylation at the same concentration as before. At 72 h posttransfection, cell culture supernatants were collected, 10-fold serially diluted, and used to inoculate MDCK cells. At 48 h postinfection, the viral titers were assessed by counting infectious centers (a group of at least five neighboring cells expressing eGFP) in live cell culture with the use of the fluorescent microscope as described above.
Assessment of HA content. The recombinant viruses were propagated in MDCK cells, concentrated, and purified via ultracentrifugation by a standard procedure. Ninety-six-well plates (Falcon-BD, Franklin Lakes, NJ) were coated with a twofold serially diluted virus preparation in PBS (pH 7.4) and incubated overnight at 4°C. The plates were washed four times in PBST (PBS containing 0.05% Tween-20), and then blocked with 1% BSA in PBS at room temperature for 1 h, followed by four washes in PBST. As a primary antibody, either a polyclonal ferret serum directed against HA (1:1,000) (32) or a mouse MAb IVC102 (1:4,000) (Biodesign, Inc., Saco, ME) was used in diluent (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.1% BSA, 0.05% Tween-20). The plates were incubated at room temperature for 1 h and then washed four times in PBST. As a secondary antibody, either a goat anti-mouse immunoglobulin G (1:3,000) (Sigma, St. Louis, MO) or a goat anti-ferret immunoglobulin G (1:3,000) (KPL, Gaithersburg, MD) conjugated with horseradish peroxidase was utilized. The plates were again incubated at room temperature for 1 h and then washed four times in PBST. The substrate TMB was added and after incubation at room temperature for 15 min, the reaction was stopped by the addition of 25 μl of 0.5 M H2SO4. The optical density was read at 405 nm with the use of Victor3 1420 multilabel counter (Elmer-Perkin, Boston, MA). Based on the optical density readings, a limiting dilution (a dilution that gives a signal equal two times the standard deviation values above the background) for each virus preparation was independently determined with the use of two different anti-HA antibodies. The virus preparation was diluted accordingly to those readings and then tested by hemagglutination assay.
Hemagglutination assay. It was shown that modern influenza A viruses efficiently agglutinate guinea pig erythrocytes but interact poorly with chicken erythrocytes (34). This property was utilized in the present study. The standardized by HA content virus preparations were first tested by HA assay with 0.5% suspension of guinea pig erythrocytes. Each virus preparation was serially twofold diluted in PBS (pH = 7.4), and then incubated with an equal volume (50 μl) of 0.5% suspension of erythrocytes in 96-well V-shaped plates (Fisher, Pittsburgh, PA) at 4°C for 1 h, and HA titers were recorded. The virus preparations were then assayed with the use of chicken erythrocytes (Rockland Immunochemicals, Gilbertville, PA) at either 4°C or 37°C. All components of the reaction were brought to an indicated temperature before combining together.
Neuraminidase inhibition assay. The NA inhibitor susceptibility of viruses was assessed by a fluorometric assay by Potier et al. (41a) with modifications as previously described (13). The IC50, which is the drug concentration needed to inhibit the enzyme activity by 50%, was evaluated by measuring NA activity in the presence of serial half-log dilutions (from 10 μM to 0.01 nM) of the NA inhibitor. After equal volumes of virus and inhibitor had been mixed and incubated at room temperature for 30 min, fluorogenic substrate [2'-(4-methylumbelliferyl)--D-N-acetylneuraminic acid; Sigma, St. Louis, MO] was added to a final concentration of 100 mM, and the reaction mixture was then incubated at 37°C for 1 h. The reaction was stopped by addition of a stop solution (150 μl of 0.5 M NaOH, pH 10.7, containing 25% ethanol), and the fluorescence was measured with the use of a Victor3 1420 multilabel counter. The excitation wavelength was 365 nm, and the emission wavelength was 460 nm.
Cell protection assay. Confluent monolayers of MDCK cells were inoculated with virus at an MOI of 0.001 TCID50 per cell, and 10-fold serial dilutions (ranging from 0.01 μM to 100 μM) of either oseltamivir carboxylate or zanamivir were added to the medium. Antiviral effect was assessed at 48 h postinfection, at which time point, control (no drug) cell monolayers infected with virus showed extensive (>90%) CPE. The cell monolayers were washed with DPBS to remove dead cells, and the number of viable cells was determined by a colorimetric method. This assay is based on the in situ reduction of 3-(4,5-dimethylthizol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) by viable cells (36). After addition of MTT to a final concentration of 1 mg/ml, the cells were incubated at 37°C for 2 h. The resulting formazan precipitate was dissolved with isopropanol containing 0.04 N HCl, and the absorbance of the solution was determined spectrophotometrically at 570 nm with the use of a Victor3 1420 multilabel counter. The antiviral effect for each virus was expressed as the EC50 (the drug concentration at which 50% of infected cells were protected against CPE).
Virus yield reduction assay. MDCK cells were infected (MOI = 0.001) and after 1 h of virus adsorption, the cells were washed with DPBS and then cultured in DMEM supplemented with 1 μg/ml TPCK-trypsin with or without zanamivir. The final concentrations of the inhibitor were 0, 0.01, 1, and 100 μM. To determine the virus yields, the supernatants from the infected cells were harvested 48 h postinfection, and titers were determined for infectivity in MDCK cells.
RESULTS
First, we generated a series of recombinant viruses that contained the HA gene from the CH/95 virus and the remaining genes from the WSN virus and differed from each other only by the number and location of the potential glycosylation sites near the receptor-binding site (Fig. 1; Table 1). The recombinant viruses were subjected to SDS-PAGE analysis, which revealed the anticipated difference in electrophoretic mobility of the HA0 molecules, depending on the number of available sites for glycan attachment (Fig. 2A). Pretreatment of the samples with an enzyme that removes N-glycans (PNG-ase F) was accompanied by a loss of the difference in the HA0 electrophoretic mobilities (Fig. 2B). Therefore, the results of the SDS-PAGE analysis confirmed that a mutation(s) introduced into the HAs of the recombinant viruses abolished glycan attachment at single, double, or triple sites (94a, 129, and 163).
To determine how the introduced changes in the HA affected the efficiency of binding to cellular receptors by the recombinant viruses, we used the host-specific hemagglutination assay (21). The CH/95 virus, as well as other contemporary human influenza A viruses (34), efficiently agglutinates guinea pig erythrocytes but not chicken erythrocytes. The virus preparations were standardized by HA content with the use of anti-HA antibodies by enzyme-linked immunosorbent assay and tested with guinea pig erythrocytes. The difference in the reciprocal HA titers was 2 fold among the virus preparations (128 to 256 hemagglutination units per 50 μl), which is in a range of this assay error. In contrast, there was a substantial difference in the HA titers determined with the use of chicken erythrocytes. The recombinant variant, whose glycans at residues 94a, 129, and 163 had been removed (Table 1; G0) efficiently agglutinated chicken erythrocytes, even at elevated temperatures (37°C). In the presence of glycan at 129, the reciprocal HA titer was reduced by 8 fold. The HA titers were even lower, when glycan at residue 163 was present alone or in combination with the other glycans. In contrast, glycan at residue 94a either did not interfere with heamagglutination (Table 1, G1) or even improved the virus ability to agglutinate chicken erythrocytes (Table 1, G1,2 versus G2).
Assessment of virus susceptibility to NA inhibitors. Because the recombinant viruses shared the NA gene (WSN), we anticipated that they would be equally susceptible to oseltamivir carboxylate by NA inhibition assay. Indeed, the estimated IC50s for the recombinant viruses (n = 8) felt into a narrow range from 0.8 nM to 2 nM (Table 1). However, the susceptibility of the recombinant viruses to NA inhibitor differed dramatically in MDCK cell culture. Based on the results of an MTT assay, the virus lacking all three glycans was highly susceptible to oseltamivir carboxylate (Table 1, G0). Peculiarly, addition of glycan at residue 94a (G1) enhanced by >4 fold the drug susceptibility of the virus in MDCK cells (0.04 μM versus <0.01 μM). In contrast, the presence of glycan at 129 was accompanied by substantial (>100-fold) reduction in the susceptibility to oseltamivir carboxylate, yet this effect was annulled by glycan at 94a (Table 1, G2 and G1,2). The viruses carrying glycan at 163, alone or in a combination with the other two glycans, were resistant to oseltamivir carboxylate (>10,000 fold). Importantly, the MTT assays conducted in the presence of the other NA inhibitor zanamivir produced similar results (Table 1).
The viruses carrying glycan at 163 were essentially zanamivir resistant, based on the results of the infectious yield reduction assay (Table 1). They exhibited a <10-fold reduction in the virus titers in the cell culture supernatants at the highest drug concentration tested (100 μM). In contrast, the viruses lacking all three glycans or carrying glycan at 94a, exhibited a >5-magnitude reduction, whereas the viruses carrying glycan at 129 exhibited an intermediate drug susceptibility (Table 1). Taken together, these results indicate that glycans attached at the HA receptor-binding site can produce the opposite effects on the virus agglutinating ability and susceptibility to NA inhibitors in MDCK cells.
Next, we assessed whether the HA amino acid backbone would influence the glycans' effect on the virus susceptibility to NA inhibitors. To this end, the HA of the WSN virus, which already contains glycan at 129, was modified to create an additional glycosylation site at residue 94a and/or 163 (Fig. 2C). The IC50s for the generated viruses assessed against oseltamivir carboxylate ranged from 2.0 nM to 3.0 nM by NA inhibition assay (Table 2). By MTT assay, the virus carrying glycans at residues 94a and 129 exhibited an approximately 10-fold increase (Table 2; EC50 = 0.2 μM) in susceptibility to zanamivir and oseltamivir carboxylate compare to the wild type virus, whereas addition of glycan at 163 resulted in drug resistance (EC50, >100 μM). The results of the yield reduction assay produced similar results (Table 2). Specifically, the virus that acquired glycan at 94a was the most susceptible to zanamivir with 10,000-fold reduction in the virus titer (Table 2, G1,2). In contrast, the virus with glycan at 163 showed the least drug susceptibility, with 4-fold reduction in the viral titers at the highest drug concentration tested (Table 2, G2,3). The negative effect of glycan at 163 on drug susceptibility was to some extent neutralized by glycan at 94a (50-fold reduction).
Therefore, we established that the presence of glycan at residue 163 caused drastic reduction of the viruses' susceptibility to NA inhibitors regardless of the HA amino acid backbone (CH/95 or WSN).
Rescue of NA-lacking virus variants. HA glycan at 163 apparently provided a compensatory mechanism for diminished NA activity in the presence of NA inhibitor. The question remained of whether this mechanism would be sufficient to permit NA independence of the virus in cell culture. To test this possibility, reverse genetics experiments were conducted, in which we attempted a rescue of virus variants that were completely devoid of NA activity due to deletion in the NA gene (32). The defective NA gene was modified to encode eGFP instead of the NA ectodomain, which allows monitoring of virus spread in live cell culture (32). At 72 h after transfection, the supernatants were harvested and used to inoculate fresh MDCK cell monolayers. The NA-lacking viruses containing glycan at residue 163 in the HA (CH/95 or WSN) were successfully rescued (Table 3). In contrast, we were unable to rescue NA-lacking viruses whose HAs were missing this glycan.
Effect of HA glycan composition on the virus requirement for NA activity. If NeuAc moieties in the vicinity of the HA receptor-binding site indeed provide a compensation of the lack of NA activity, a carbohydrate structure of glycan at 163 should be critical for release of NA-lacking viruses. Glycoproteins are initially glycosylated by transfer of Glc3Man9GlcNAc2 from a lipid donor to Asn of the protein. The oligosaccharide is subsequently processed to the mature forms containing five to nine mannose residues (high-mannosidic type). Further trimming and elongation reactions in the Golgi apparatus result in formation of complex-type oligosaccharides. Trimming reactions proceed through the action of endomannosidase and Golgi mannosidases I and II. Related and depending on trimming, elongation by the families of N-acetylglucosaminyl-, fucosyl-, galactosyl-, and sialyltransferases takes place (43). dMj, an inhibitor of mannosidase I, blocks formation of complex glycans, resulting in the accumulation of high mannose oligosaccharides with a predominant species being Man9GlcNAc2 (8). Swainsonine, an inhibitor of mannosidase II, also blocks formation of complex glycans and causes accumulation of hybrid-type glycans (17, 30).
MDCK cells were infected with the recombinant G1,2,3 variant and, immediately after virus adsorption, either dMj or swainsonine was added to the medium. At 4 h postinfection, virus proteins were metabolically labeled with [35S]methionine-cysteine in the presence of either dMJ or swainsonine in the medium. The proteins incorporated into the virus particles were analyzed by SDS-PAGE (Fig. 3). The slowest-migrating band was produced by the HA0 from the cells maintained in the absence of N-glycosylation inhibitors (Fig. 3, lane 1), whereas the fastest-moving band was produced by the HA0 synthesized in the presence of dMj (Fig. 3, lane 4). Hydrolysis with endo H produced a substantial shift in the bands synthesized in the presence of inhibitors of N-glycosylation (Fig. 3, lanes 6 and 9), whereas it had a much lesser effect on the electrophoretic mobility of the HA0 synthesized in the absence of the inhibitors (Fig. 3, lane 3). Hydrolysis with PNGase F resulted in a single intense band of the same size that was detected in all three immunoprecipitates.
Having confirmed that inhibitors of N-glycosylation (dMj and swainsonine) arrest the HA glycosylation at different steps, we next tested their effects on the outcome of the rescue experiments. When administered to cell culture medium shortly after transfection, neither inhibitor significantly affected the rescue and the infectious yields of the generated virus with the wild-type NA (Table 4). Arrest of the HA glycosylation at a hybrid step by swainsonine also had no apparent effect on the rescue and yields of the NA activity-lacking virus carrying the wild-type HA and caused a 4- to 25-fold reduction in the quantity of the NA activity-lacking virus carrying a single glycan at 163 (Table 4). In contrast, dMj, which arrests the HA glycosylation at the high-mannosidic step, caused a drastic reduction in the yields of the NA activity-lacking viruses, especially of the virus with the single glycan at 163 (50 TCID50/ml).
DISCUSSION
A balance between the antagonistic HA and NA functions can be altered by various events, such as virus transmission to a new host, reassortment, or therapeutic intervention (22, 33, 49). It has been reported that both HA and NA determine the sensitivity of naturally occurring avian influenza viruses to zanamivir in vitro (1, 2). The results of the present study present direct evidence for the role of the HA glycosylation in the human influenza virus susceptibility to NA inhibitors and thus provide a further insight into the functional balance of HA and NA. Because the yields of the NA-lacking mutants in the supernatants were attenuated by 10 to 100 fold, compared to those of the viruses' wild-type NA, HA glycosylation provided only a partial compensation for the loss of the NA activity. Both the oligosaccharide composition and attachment site of HA glycans affected the virus requirement for NA activity in cell culture. Glycosylation at residue 163 on the tip of the HA molecule provided a compensatory mechanism for the loss of NA activity which appears to stem from the reduced efficiency of the HA binding to cellular receptors (2, 38). In MDCK cells, desialidation of the HA by viral NA occurs on the cell surface (26). Therefore, when virus is grown in the presence of NA inhibitor, only complex and hybrid-type HA glycans would retain the terminal NeuAc moieties, whereas the structure of high-mannose-type glycans should not be affected (7). The HA1 subunit of the contemporary CH/95 strain contains eight potential sites of glycosylation at Asn 21, 33, 63, 94a, 129, 163, 271, and 289 (H3 numbering) (GenBank accession no. AF398878) (52). Some of those glycans appear to be essential for HA functioning (48), and the corresponding glycosylation motifs are conserved in the majority of strains. The A/WSN/33 virus underwent intensive adaptation to mice and various cell cultures; its HA1 glycosylation pattern differs from that of the other viruses: four potential sites of glycosylation are present at Asn 21, 65, 129, and 271 (18). Information on glycan composition for the human influenza virus HA is sparse (31, 37), yet the available data correlate well with our findings. Glycan at 94a is mostly likely to be of a high-mannosidic type (15), whereas glycan attached at 129, at least for WSN virus grown in MDBK cells, belongs to a complex type (31). We demonstrated that glycans attached at 129 and 163 reduced the agglutinating activity toward chicken erythrocytes, with glycan at 163 playing a dominant role. Importantly, when propagated in the presence of NA inhibitor, the virus carrying the glycan 163 and wild-type NA failed to agglutinate chicken erythrocytes at 4°C (results not shown), indicating the role of terminal neuraminic acid moieties. This result is in accordance with the results of the hemadsorption studies on the cell surface-expressed avian HA (38). To abolish a glycosylation motif, Asn 163 was replaced by Gln in the present study. Since this residue is not a part of the receptor-binding site (27), we do not believe that amino acid substitution by itself was responsible for the observed difference in the virus properties. Glycan at 94a, when present alone or in combination with glycans 129 and 163, enhanced virus susceptibility to NA inhibitors (Tables 1 and 2). The molecular mechanism for this differential effect is not clear. For example, glycan at Asn 94a could enhance HA binding efficiency via auxiliary interaction with a high-mannose specific lectin expressed on the cell surface (i.e., VIP36 of MDCK cells) (14).
We addressed the question of importance of oligosaccharide chain composition in compensating for the loss of the NA activity in the reverse genetics rescue experiments, when rescue of the NA activity-lacking viruses was attempted in the presence of N-glycosylation inhibitors. Swainsonine, which arrests N-glycosylation at the hybrid step, did not substantially affect the rescue of the NA activity-lacking viruses. In contrast, either we were unable to rescue NA-lacking virus or its titer was significantly reduced in the presence of dMj (Table 4). This could be explained by the ability of hybrid type oligosaccharides to retain the terminal NeuAc acid moieties. The other, less-likely explanation could be the difference in a size between a hybrid and a high-mannosidic oligosaccharide chain. We used dMj at a concentration of 15 μg/ml, which efficiently arrested N-glycosylation at a high-mannosidic step (Fig. 3); however, it may be insufficient to completely prevent formation of complex glycans (9). It is unlikely that the negative effect of dMj on the rescue of the NA-lacking mutant was due to general toxicity or reduction of viral infectivity, since no titer reduction was observed for the virus with wild-type NA in the present study (Table 4) or even when MDCK cells were cultured at much greater dMj concentrations (1 mM) (D.F. Smee, personal communication). Acquisition of glycans at 129 and 163 at the receptor-binding site (as well as those at residues 131 and 158 in some viruses) is believed to be a result of antigenic drift of the A/H1N1 viruses in the human population. Masking antigenic epitopes by glycans is one of the mechanisms by which virus escapes neutralization by preexisting antibodies (45). It is believed that a size of glycan is not critical for epitope masking (35). In contrast, glycans containing terminal NeuAc moieties should contribute the most to the virus's ability to escape the antiviral effect of NA inhibitors. Prior studies have found that the recombinant WSN virus carrying the HA and NA of the 1918 virus exhibited oseltamivir susceptibility in MDCK cells (47). Of note, this virus (42) had a glycosylation site at 94a but lacked sites at position 129 and 163. Such viruses therefore should be more susceptible to NA inhibitors than the modern human viruses of the H1N1 subtype, at least in MDCK cells.
In our study, glycan at 163 produced a compensatory effect, regardless of the HA amino acid sequences (WSN or CH/95). Because the HA sequences of these two viruses differed substantially due to antigenic drift, our data indicated that glycans attached at the tip of the HA molecule have the potential to affect susceptibility of a broad spectrum of influenza viruses, even beyond that of the H1 antigenic subtype. Our findings warrant a more detailed analysis of the HA glycosylation effect on human virus susceptibility to NA inhibitors.
ACKNOWLEDGMENTS
This work was supported by U.S. Public Heath Service research grant AI45782 from the National Institute of Allergy and Infectious Diseases.
We are grateful to E. Hoffmann and R. G. Webster, St. Jude Children's Research Hospital, for providing a set of eight plasmids encoding the genome of A/WSN/33 (H1N1) virus, which we used in the reverse genetic experiments.
REFERENCES
Baigent, S. J., R. C. Bethell, and J. W. McCauley. 1999. Genetic analysis reveals that both haemagglutinin and neuraminidase determine the sensitivity of naturally occurring avian influenza viruses to zanamivir in vitro. Virology 263:323-338.
Baigent, S. J., and J. W. McCauley. 2001. Glycosylation of haemagglutinin and stalk-length of neuraminidase combine to regulate the growth of avian influenza viruses in tissue culture. Virus Res. 79:177-185.
Basak, S., M. Tomana, and R. W. Compans. 1985. Sialic acid is incorporated into influenza hemagglutinin glycoproteins in the absence of viral neuraminidase. Virus Res. 2:61-68.
Baum, L. G., and J. C. Paulson. 1991. The N2 neuraminidase of human influenza virus has acquired a substrate specificity complementary to the hemagglutinin receptor specificity. Virology 180:10-15.
Chu, V. C., and G. R. Whittaker. 2004. Influenza virus entry and infection require host cell N-linked glycoprotein. Proc. Natl. Acad. Sci. USA 101:18153-18158.
Colman, P. M. 1989. Neuraminidase: enzyme and antigen, p. 175-210. In R. M. Krug (ed.), The influenza viruses. Plenum Press, New York, N.Y.
Elbein, A. D. 1991. Glycosidase inhibitors: inhibitors of N-linked oligosaccharide processing. FASEB J. 5:3055-3063.
Elbein, A. D., G. Legler, A. Tlusty, W. McDowell, and R. Schwarz. 1984. The effect of deoxymannojirimycin on the processing of the influenza viral glycoproteins. Arch. Biochem. Biophys. 235:579-588.
Elbein, A. D., J. E. Tropea, M. Mitchell, and G. P. Kaushal. 1990. Kifunensine, a potent inhibitor of the glycoprotein processing mannosidase I. J. Biol. Chem. 265:15599-15605.
Gimsa, U., I. Grotzinger, and J. Gimsa. 1996. Two evolutionary strategies of influenza viruses to escape host non-specific inhibitors: alteration of hemagglutinin or neuraminidase specificity. Virus Res. 42:127-135.
Gubareva, L. V., R. Bethell, G. J. Hart, K. G. Murti, C. R. Penn, and R. G. Webster. 1996. Characterization of mutants of influenza A virus selected with the neuraminidase inhibitor 4-guanidino-Neu5Ac2en. J. Virol. 70:1818-1827.
Gubareva, L. V., M. S. Nedyalkova, D. V. Novikov, K. G. Murti, E. Hoffmann, and F. G. Hayden. 2002. A release-competent influenza A virus mutant lacking the coding capacity for the neuraminidase active site. J. Gen. Virol. 83:2683-2692.
Gubareva, L. V., R. G. Webster, and F. G. Hayden. 2002. Detection of influenza virus resistance to neuraminidase inhibitors by an enzyme inhibition assay. Antiviral Res. 53:47-61.
Hara-Kuge, S., T. Ohkura, H. Ideo, O. Shimada, S. Atsumi, and K. Yamashita. 2002. Involvement of VIP36 in intracellular transport and secretion of glycoproteins in polarized Madin-Darby canine kidney (MDCK) cells. J. Biol. Chem. 277:16332-16339.
Hartley, C. A., D. C. Jackson, and E. M. Anders. 1992. Two distinct serum mannose-binding lectins function as beta inhibitors of influenza virus: identification of bovine serum beta inhibitor as conglutinin. J. Virol. 66:4358-4363.
Hierholzer, J. C., and R. A. Killington. 1996. Virus isolation and quantitation, p. 25-44. In B. W. J. Mahy and H. O. Kangro (ed.), Virology methods manual. Academic Press, London, United Kingdom.
Ho, J. J., R. S. Jaituni, S. C. Crawley, S. C. Yang, J. R. Gum, and Y. S. Kim. 2003. N-Glycosylation is required for the surface localization of MUC17 mucin. Int. J. Oncol. 23:585-592.
Hoffmann, E., G. Neumann, Y. Kawaoka, G. Hobom, and R. G. Webster. 2000. A DNA transfection system for generation of influenza A virus from eight plasmids. Proc. Natl. Acad. Sci. USA 97:6108-6113.
Hoffmann, E., and R. G. Webster. 2000. Unidirectional RNA polymerase I-polymerase II transcription system for the generation of influenza A virus from eight plasmids. J. Gen. Virol. 81:2843-2847.
Inkster, M. D., V. S. Hinshaw, and I. T. Schulze. 1993. The hemagglutinins of duck and human H1 influenza viruses differ in sequence conservation and in glycosylation. J. Virol. 67:7436-7443.
Ito, T., Y. Suzuki, L. Mitnaul, A. Vines, H. Kida, and Y. Kawaoka. 1997. Receptor specificity of influenza A viruses correlates with the agglutination of erythrocytes from different animal species. Virology 227:493-499.
Kaverin, N. V., A. S. Gambaryan, N. V. Bovin, I. A. Rudneva, A. A. Shilov, O. M. Khodova, N. L. Varich, B. V. Sinitsin, N. V. Makarova, and E. A. Kropotkina. 1998. Postreassortment changes in influenza A virus hemagglutinin restoring HA-NA functional match. Virology 244:315-321.
Kaverin, N. V., M. N. Matrosovich, A. S. Gambaryan, I. A. Rudneva, A. A. Shilov, N. L. Varich, N. V. Makarova, E. A. Kropotkina, and B. V. Sinitsin. 2000. Intergenic HA-NA interactions in influenza A virus: postreassortment substitutions of charged amino acid in the hemagglutinin of different subtypes. Virus Res. 66:123-129.
Kim, C. U., X. Chen, and D. B. Mendel. 1999. Neuraminidase inhibitors as anti-influenza virus agents. Antivir. Chem. Chemother. 10:141-154.
Liu, C., M. C. Eichelberger, R. W. Compans, and G. M. Air. 1995. Influenza type A virus neuraminidase does not play a role in viral entry, replication, assembly, or budding. J. Virol. 69:1099-1106.
Luo, C., E. Nobusawa, and K. Nakajima. 2002. Analysis of the desialidation process of the haemagglutinin protein of influenza B virus: the host-dependent desialidation step. J. Gen. Virol. 83:1729-1734.
Matrosovich, M., A. Tuzikov, N. Bovin, A. Gambaryan, A. Klimov, M. R. Castrucci, I. Donatelli, and Y. Kawaoka. 2000. Early alterations of the receptor-binding properties of H1, H2, and H3 avian influenza virus hemagglutinins after their introduction into mammals. J. Virol. 74:8502-8512.
McKimm-Breschkin, J. L. 2000. Resistance of influenza viruses to neuraminidase inhibitors. Antiviral Res. 47:1-17.
McKimm-Breschkin, J. L., T. J. Blick, A. Sahasrabudhe, T. Tiong, D. Marshall, G. J. Hart, R. C. Bethell, and C. R. Penn. 1996. Generation and characterization of variants of NWS/G70c influenza virus after in vitro passage in 4-amino-Neu5Ac2en and 4-guanidino-Neu5Ac2en. Antimicrob. Agents Chemother. 40:40-46.
Merkle, R. K., A. D. Elbein, and A. Heifetz. 1985. The effect of swainsonine and castanospermine on the sulfation of the oligosaccharide chains of N-linked glycoproteins. J. Biol. Chem. 260:1083-1089.
Mir-Shekari, S. Y., D. A. Ashford, D. J. Harvey, R. A. Dwek, and I. T. Schulze. 1997. The glycosylation of the influenza A virus hemagglutinin by mammalian cells. A site-specific study. J. Biol. Chem. 272:4027-4036.
Mishin, V. P., M. S. Nedyalkova, F. G. Hayden, and L. V. Gubareva. 2005. Protection afforded by intranasal immunization with the neuraminidase-lacking mutant of influenza A virus in a ferret model. Vaccine 23:2922-2927.
Mitnaul, L., M. N. Matrosovich, M. R. Castrucci, A. B. Tuzikov, N. V. Bovin, D. Kobasa, and Y. Kawaoka. 2000. Balances hemagglutinin and neuraminidase activities are critical for efficient replication of influenza A virus. J. Virol. 74:6015-6020.
Morishita, T., E. Nobusawa, K. Nakajima, and S. Nakajima. 1996. Studies on the molecular basis for loss of the ability of recent influenza A (H1N1) virus strains to agglutinate chicken erythrocytes. J. Gen. Virol. 77:2499-2506.
Munk, K., E. Pritzer, E. Kretzschmar, B. Gutte, W. Garten, and H. D. Klenk. 1992. Carbohydrate masking of an antigenic epitope of influenza virus haemagglutinin independent of oligosaccharide size. Glycobiology 2:233-240.
Nagai, T., Y. Miyaichi, T. Tomimori, Y. Suzuki, and H. Yamada. 1992. In vivo anti-influenza virus activity of plant flavonoids possessing inhibitory activity for influenza virus sialidase. Antiviral Res. 19:207-217.
Nakamura, K., A. S. Brown, and R. W. Compans. 1980. Glycosylation sites of influenza viral glycoproteins. Tryptic glycopeptides from the A/WSN (H0N1) hemagglutinin glycoprotein. Virology 107:208-221.
Ohuchi, M., R. Ohuchi, A. Fieldmann, and H. D. Klenk. 1997. Regulation of receptor binding affinity of influenza virus hemagglutinin by its carbohydrate moiety. J. Virol. 71:8377-8384.
Palese, P., and R. W. Compans. 1976. Inhibition of influenza virus replication in tissue culture by 2-deoxy-2,3-dehydro-N-trifluoro-acetyl-neuraminic acid (FANA): mechanism of action. J. Gen. Virol. 33:159-163.
Pan, Y. T., H. Hori, and A. D. Elbein. 1987. The effect of glycoprotein-processing inhibitors on the secretion of glycoproteins by Madin-Darby canine kidney cells. Biochem. Cell Biol. 65:345-353.
Penn, C. R., L. M. Barnett, R. C. Bethell, R. Fenton, K. L. Gearing, N. Healy, and A. J. Jowett. 1996. Selection of influenza virus with reduced sensitivity in vitro to the neuraminidase inhibitor GG167 (4-guanidino-Neu5Ac2en): changes in the haemagglutinin may compensate for loss of neuraminidase activity, p. 735-740. In L. E. Brown, A. W. Hampson, and R. G. Webster (ed.), Options for the control of influenza III. Elsevier Science B.V., Amsterdam, The Netherlands.
Potier, M., L. Mameli, M. Belisle, L. Dallaire, and S. B. Malencon. 1979. Fluorometric assay of neuraminidase with a sodium (4-methylumbelliferyl-alpha-D-N-acetylneuraminate) substrate. Anal. Biochem. 94:287-296.
Reid, A. H., T. G. Fanning, J. V. Hultin, and J. K. Taubenberger. 1999. Origin and evolution of the 1918 "Spanish" influenza virus hemagglutinin gene. Proc. Natl. Acad. Sci. USA 96:1651-1666.
Roth, J. 2002. Protein N-glycosylation along the secretory pathway: relationship to organelle topography and function, protein quality control, and cell interactions. Chem. Rev. 102:285-303.
Shinya, K., Y. Fujii, H. Ito, T. Ito, and Y. Kawaoka. 2004. Characterization of a neuraminidase-deficient influenza a virus as a potential gene delivery vector and a live vaccine. J. Virol. 78:3083-3088.
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.
Stray, S. J., R. D. Cummings, and G. M. Air. 2000. Influenza virus infection of desialylated cells. Glycobiology 10:649-658.
Tumpey, T. M., A. Garcia-Sastre, A. Mikulasova, J. K. Taubenberger, D. E. Swayne, P. Palese, and C. F. Basler. 2002. Existing antivirals are effective against influenza viruses with genes from the 1918 pandemic virus. Proc. Natl. Acad. Sci. USA 99:13849-13854.
Wagner, R., D. Heuer, T. Wolff, A. Herwig, and H. D. Klenk. 2002. N-Glycans attached to the stem domain of haemagglutinin efficiently regulate influenza A virus replication. J. Gen. Virol. 83:3-9.
Wagner, R., M. Matrosovich, and H. D. Klenk. 2002. Functional balance between haemagglutinin and neuraminidase in influenza virus infections. Rev. Med. Virol. 12:159-166.
Wagner, R., T. Wolff, A. Herwig, S. Pleschka, and H. D. Klenk. 2000. Interdependence of hemagglutinin glycosylation and neuraminidase as regulators of influenza virus growth: a study by reverse genetics. J. Virol. 74:6316-6323.
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.
Winter, G., S. Fields, and G. G. Brownlee. 1981. Nucleotide sequence of the haemagglutinin gene of a human influenza virus H1 subtype. Nature 292:72-75.(Vasiliy P. Mishin, Dmitri)
Department of Pathology, School of Medicine, University of Virginia, Charlottesville, Virginia
ABSTRACT
Inhibition of neuraminidase (NA) activity prevents release of progeny virions from influenza-infected cells and removal of neuraminic (sialic) acid moieties from glycans attached to hemagglutinin (HA). Neuraminic acid moieties situated near the HA receptor-binding site can reduce the efficiency of virus binding and decrease viral dependence on NA activity for replication. With the use of reverse genetics technique, we investigated the effect of glycans attached at Asn 94a, 129, and 163 on the virus susceptibility to NA inhibitors in MDCK cells and demonstrated that the glycan attached at Asn 163 plays a dominant role in compensation for the loss of NA activity.
INTRODUCTION
The two surface glycoproteins, hemagglutinin (HA) and neuraminidase (NA), of influenza A virus interact with cellular receptors containing terminal neuraminic acid (NeuAc) moieties. HA initiates infection by binding to cellular receptors, whereas NA destroys the receptors by cleaving off NeuAc moieties (49). In addition, HA binds to complementary, as-yet-unidentified, cellular receptors that lack NeuAc moieties (5, 46). The functional balance between HA and NA in influenza virus infections has been intensively studied (4, 10, 22, 23, 33, 49, 50). The HA precursor (HA0) is proteolytically cleaved into two subunits, HA1 and HA2; the HA1 subunit carries the NeuAc-binding site, and the HA2 subunit is responsible for fusion of viral and cellular membranes (51). NA is not essential for influenza A virus assembly or budding (25), but its enzymatic activity facilitates the progeny virions' release from infected cells (6). The anti-influenza drugs oseltamivir and zanamivir inhibit NA activity by targeting the enzyme active site formed by highly conserved residues (6, 24). Substitutions at those residues confer virus resistance to NA inhibitors (28). However, drug resistance can also be conferred via an NA-independent mechanism. Amino acid substitutions in HA can lessen viral dependence on NA activity for release from infected cells and thus decrease susceptibility to NA inhibitors by reducing efficiency of virus binding to cellular receptors (11, 29, 41, 49).
Propagation of virus in the presence of an NA inhibitor should result in retention of NeuAc moieties by the complex glycans attached to HA (3, 39). Of note, negatively charged NeuAc moieties situated near the HA receptor-binding site have a potential to impair the virus binding to cellular receptors (38). Therefore, viruses containing complex glycans at the receptor-binding site would exhibit lower susceptibility to NA inhibitors than those lacking such glycans. To investigate this possibility, we utilized the influenza A viruses that belong to the H1N1 antigenic subtype as a model. Viruses of this subtype produced the devastating "Spanish flu" pandemic in 1918 (42), and antigenic drift variants of this virus remained in circulation in the human population until mid-1950s. In 1977, the virus reemerged in the human population, causing the "Russian flu" epidemic; since that time, antigenic drift variants have been in circulation in the human population. There is a considerable variance in the glycosylation patterns near the HA receptor-binding site among the human influenza A/H1N1 viruses, due to antigenic drift and host adaptation (20). For example, the so-called early laboratory-passaged virus A/WSN/33 (WSN) contains a single glycosylation site at Asn129, whereas the contemporary virus, A/Charlottesville/31/95 (CH/95), contains two additional glycosylation sites at Asn 94a and Asn 163) (Fig. 1) (numbering according to reference 52). In our previous studies, we demonstrated that both viruses were equally susceptible to NA inhibitors by the enzyme inhibition assay, whereas in cell culture, the CH/95 virus was drug resistant and the WSN virus was drug susceptible (12). Moreover, reassortant virus that contained the HA gene from the CH/95 virus in the WSN virus genetic background also exhibited drug resistance in cell culture (12). In the present study, we investigated the contribution of HA glycans attached at Asn 94a, 129, and 163 to viral dependence on NA activity. Our findings provide further insights into molecular mechanisms underlying the human virus resistance to NA inhibitors in cell culture.
MATERIALS AND METHODS
Compounds. Inhibitors of influenza virus neuraminidase activity zanamivir (GlaxoWellcome Research and Development, Stevenage, United Kingdom) and oseltamivir (Roche Laboratories, Inc., Nutley, NJ) were provided by their respective manufacturers. Inhibitors of N-glycosylation 1-deoxymannojirimycin (dMj) and swainsonine were purchased from Sigma, St. Louis, MO.
Viruses and cells. The A/Charlottesville/31/95 (H1N1) (CH/95) virus was from the repository of the Respiratory Disease Study Unit at the University of Virginia. The recombinant A/WSN/33 (H1N1) virus and its reassortants containing the HA gene from the CH/95 virus were generated by the reverse genetics approach (see below). Virus stocks prepared in Madin-Darby canine kidney (MDCK) cells were stored at –80°C. MDCK cells (a kind gift of Alan Hay, Mill Hill, United Kingdom) were utilized in all experiments, including determinations of virus titer and assessment of drug susceptibility. MDCK cells were grown in 24-well plates in Eagle's minimum essential medium (EMEM) (Cambrex, Wallorsville, MD) supplemented with 10% fetal calf serum (HyClone, Logan, UT) and antibiotics. Before infection, cell monolayers were rinsed twice with Dulbecco's phosphate-buffered saline (DPBS) to remove serum. Virus was inoculated onto confluent cellular monolayers and after 1 h of adsorption at room temperature, the cells were washed with DPBS to remove unbound virus. EMEM supplemented with N-tosyl-L-phenylalanine chloromethyl ketone (TPCK)-treated trypsin (Worthington, Lakewood, NJ) to a final concentration of 1 μg/ml was added to cells. The cells were incubated at 37°C for 48 h, at which point cell culture supernatants were harvested and used to assess virus titers in MDCK cells by a standard assay (16).
Plasmids construction. With the use of PCR-directed mutagenesis and high-fidelity KOD DNA polymerase (Novagen, Madison, WI), plasmid pHW-HA31 encoding the HA gene of CH/95 virus (12) was modified to abolish a glycosylation motif (Asn-X-Thr/Ser) at residues 94a, 129, and 163. To remove a glycosylation site at residue 94a, the sequence AAT-GGA-ACA (Asn-Gly-Thr) was changed to CAG-GGA-ACA (Gln-Gly-Thr). To remove sites at position 129, the sequence AAC-CAC-ACC (Asn-His-Thr) was changed to CAG-CAC-ACC (Gln-His-Thr), and at residue 163, AAT-CTG-AGC (Asn-Leu-Ser) was changed to CAG-CTG-AGC (Gln-Leu-Ser) (Table 1). In addition, plasmid pHW184 (18) encoding the HA gene (WSN) was modified to create a glycan attachment site at residue 94a and/or 163. The sequence AAT-GGA-GCA (94aAsn-Gly-Ala) was changed to AAT-GGA-ACA (94aAsn-Gly-Thr) and/or the sequence AAG-CTG-ACC (163Lys-Leu-Thr) was changed to AAT-CTG-AAC (163Asn-Leu-Thr). Lack of unwanted mutations was confirmed by sequence analysis of the plasmid DNAs.
Sequence analysis. Extraction of viral RNA from MDCK cell supernatants was performed with the use of RNAeasy kit according to the manufacturer's manual (QIAGEN, Clarita, CA). The primer U9 5'-AGCAGAAGC-3' was utilized to generate cDNA with the use of MMLV reverse transcriptase (Promega, Madison, WI). The HA gene was then amplified with a use of a standard PCR and the PCR product was then purified with the use of the QIAquick PCR purification kit according to the manufacturer's instructions (QIAGEN, Clarita, CA). Purified PCR product was subjected to sequence analysis with the use of ABI 373 DNA sequencer (Applied Biosystems, Foster City, CA) at the Center of Biotechnology at the University of Virginia. Sequencher 4.0 software (Gene Codes Corporation, Ann Arbor, MI) was used for analysis and translation of nucleotide sequence data.
Procedure for generation of HA glycosylation site variants. To generate recombinant viruses, the eight-plasmid reverse genetics system that allows synthesis of both templates vRNA (negative sense) and mRNA (positive sense) from the same plasmid was utilized (19). Coculture of 293T human embryonic kidney cells and MDCK cells was maintained in six-well plates in EMEM supplemented with 10% fetal calf serum. Cells were washed with DPBS, and then OptiMEM medium (Invitrogen, Carlsbad, CA) supplemented with antibiotics and antimycotics was added. Cells were then transfected with a mixture of eight plasmid DNAs (total of 1 μg/well) by using TransIt transfection reagent (Panvera, Madison, WI) as previously described (18). When specified, plasmid encoding a modified HA gene (either WSN or CH/95) was used instead of plasmid encoding the wild-type HA (WSN). At 24 h posttransfection, cell culture supernatant was replaced with EMEM supplemented with 0.3% bovine serum albumin (BSA; Gibco, Carlsbad, CA) and TPCK-treated trypsin at a final concentration of 1 μg/ml. Supernatants containing the recombinant viruses were collected 48 h later and used to infect MDCK cell monolayers by a standard procedure. Lack of unwanted mutations in HA genes of recombinant viruses was confirmed by the sequence analysis.
Radioactive labeling of viral proteins and sodium dodecyl sulfate-polyacrylamide gel electrophoresis (SDS-PAGE) analysis. MDCK cell monolayers were infected with virus at a multiplicity of infection (MOI) of 5 50% tissue culture infective doses (TCID50) per cell. At 4 h postinfection, the cells were metabolically radiolabeled with [35S]methionine-cysteine (100 μCi per 106 cells) (Perkin-Elmer, Boston, MA). A 15-min pulse was followed by a 45-min chase, and then the supernatants were discarded and the cells were lysed in RIPA buffer (50 mM Tris-HCl, pH 7.5, 150 mM NaCl, 1% IGEPAL CA-630, 0.5% sodium deoxycholate, 0.1% SDS, and 2 mM EDTA). The precursor HA0 protein was immunoprecipitated with anti-HA monoclonal antibodies (MAb) IVC102. To remove N-glycosylated moieties, an aliquot of each immunoprecipitate was pretreated with of peptide:N-glycosidase F (PNGase F; New England Biolabs, Beverly, MA) according to the manufacturer's recommendations for 6 h at 37°C.
To analyze the effect of inhibitors of N-glycosylation on the oligosaccharide composition of the HA molecules incorporated into the virus particles, MDCK cells were infected at an MOI of 5; after 1 h of virus adsorption, EMEM containing either dMj (Sigma, St. Louis, MO) or swainsonine (Sigma, St. Louis, MO) at a final concentration of 75 μM (15 μg/ml) and 60 μM (10 μg/ml), respectively, was added. At 4 h postinfection, the supernatants were removed and the cells were metabolically radiolabeled with [35S]methionine-cysteine (100 μCi per 106) in the presence of either dMj or swainsonine. At 18 h postinfection, the cell culture supernatants were harvested, and the cellular debris was removed by centrifugation at 12,000 rpm in an Eppendorf centrifuge for 15 min and then subjected to ultracentrifugation through a 25% sucrose cushion at 27,000 rpm (Beckman centrifuge, rotor SW28) for 90 min. The pellet containing the virus particles was resuspended in 50 μl of a loading gel buffer containing SDS. The denatured proteins were then incubated with either endo H glycosidase (500 U; New England Biolabs, Beverly, MA) or PNGase F for 12 h. At the end of the digestions, the samples and the control incubations without enzyme were examined by 8% SDS-PAGE, followed by fluorography.
Rescue procedure for recombinant NA activity-lacking virus variants. Generation of viruses lacking NA activity was done by transfection of 293T cells cocultured with MDCK cells, essentially as described above. The plasmid encoding the wild-type NA (WSN) was replaced with the plasmid (pdelNA/eGFP) that encodes the enhanced green fluorescence protein (eGFP) instead of the NA active site (32, 44). Virus spread was monitored in live cells with the use of Zeiss Axiovert 135 TV inverted microscope equipped with a cooled charge-coupled device camera (Hamamatsu Orca, Japan) and with excitation (488 nm) and emission (530 nm) filters (ChromaTechnology Corp., Brattleboro, VT). At 72 h posttransfection (48 h after addition of trypsin to the medium), the supernatants were harvested and used to inoculate fresh monolayers of MDCK cells. Rescue was considered successful if cytopathic effect (CPE) was observed on day 3 postinfection and HA activity was detected in MDCK cell supernatants.
Rescue of NA-lacking virus variants in the presence of inhibitors of N-glycosylation. Three hours after transfection, cell culture supernatants were removed and cells were washed with DPBS. OptiMEM containing either dMj (Sigma, St. Louis, MO) at a final concentration of 75 μM (15 μg/ml) or swainsonine (Sigma, St. Louis, MO) at a final concentration of 60 μM (10 μg/ml) was added. These concentrations were reported to produce desirable effects on composition of oligosaccharide chains of glycoproteins synthesized in various cell cultures including MDCK cells (8, 9, 17, 40). At 24 h posttransfection, supernatants were replaced with fresh EMEM containing TPCK-trypsin and the corresponding inhibitor of N-glycosylation at the same concentration as before. At 72 h posttransfection, cell culture supernatants were collected, 10-fold serially diluted, and used to inoculate MDCK cells. At 48 h postinfection, the viral titers were assessed by counting infectious centers (a group of at least five neighboring cells expressing eGFP) in live cell culture with the use of the fluorescent microscope as described above.
Assessment of HA content. The recombinant viruses were propagated in MDCK cells, concentrated, and purified via ultracentrifugation by a standard procedure. Ninety-six-well plates (Falcon-BD, Franklin Lakes, NJ) were coated with a twofold serially diluted virus preparation in PBS (pH 7.4) and incubated overnight at 4°C. The plates were washed four times in PBST (PBS containing 0.05% Tween-20), and then blocked with 1% BSA in PBS at room temperature for 1 h, followed by four washes in PBST. As a primary antibody, either a polyclonal ferret serum directed against HA (1:1,000) (32) or a mouse MAb IVC102 (1:4,000) (Biodesign, Inc., Saco, ME) was used in diluent (20 mM Tris-HCl, pH 7.5, 150 mM NaCl, 0.1% BSA, 0.05% Tween-20). The plates were incubated at room temperature for 1 h and then washed four times in PBST. As a secondary antibody, either a goat anti-mouse immunoglobulin G (1:3,000) (Sigma, St. Louis, MO) or a goat anti-ferret immunoglobulin G (1:3,000) (KPL, Gaithersburg, MD) conjugated with horseradish peroxidase was utilized. The plates were again incubated at room temperature for 1 h and then washed four times in PBST. The substrate TMB was added and after incubation at room temperature for 15 min, the reaction was stopped by the addition of 25 μl of 0.5 M H2SO4. The optical density was read at 405 nm with the use of Victor3 1420 multilabel counter (Elmer-Perkin, Boston, MA). Based on the optical density readings, a limiting dilution (a dilution that gives a signal equal two times the standard deviation values above the background) for each virus preparation was independently determined with the use of two different anti-HA antibodies. The virus preparation was diluted accordingly to those readings and then tested by hemagglutination assay.
Hemagglutination assay. It was shown that modern influenza A viruses efficiently agglutinate guinea pig erythrocytes but interact poorly with chicken erythrocytes (34). This property was utilized in the present study. The standardized by HA content virus preparations were first tested by HA assay with 0.5% suspension of guinea pig erythrocytes. Each virus preparation was serially twofold diluted in PBS (pH = 7.4), and then incubated with an equal volume (50 μl) of 0.5% suspension of erythrocytes in 96-well V-shaped plates (Fisher, Pittsburgh, PA) at 4°C for 1 h, and HA titers were recorded. The virus preparations were then assayed with the use of chicken erythrocytes (Rockland Immunochemicals, Gilbertville, PA) at either 4°C or 37°C. All components of the reaction were brought to an indicated temperature before combining together.
Neuraminidase inhibition assay. The NA inhibitor susceptibility of viruses was assessed by a fluorometric assay by Potier et al. (41a) with modifications as previously described (13). The IC50, which is the drug concentration needed to inhibit the enzyme activity by 50%, was evaluated by measuring NA activity in the presence of serial half-log dilutions (from 10 μM to 0.01 nM) of the NA inhibitor. After equal volumes of virus and inhibitor had been mixed and incubated at room temperature for 30 min, fluorogenic substrate [2'-(4-methylumbelliferyl)--D-N-acetylneuraminic acid; Sigma, St. Louis, MO] was added to a final concentration of 100 mM, and the reaction mixture was then incubated at 37°C for 1 h. The reaction was stopped by addition of a stop solution (150 μl of 0.5 M NaOH, pH 10.7, containing 25% ethanol), and the fluorescence was measured with the use of a Victor3 1420 multilabel counter. The excitation wavelength was 365 nm, and the emission wavelength was 460 nm.
Cell protection assay. Confluent monolayers of MDCK cells were inoculated with virus at an MOI of 0.001 TCID50 per cell, and 10-fold serial dilutions (ranging from 0.01 μM to 100 μM) of either oseltamivir carboxylate or zanamivir were added to the medium. Antiviral effect was assessed at 48 h postinfection, at which time point, control (no drug) cell monolayers infected with virus showed extensive (>90%) CPE. The cell monolayers were washed with DPBS to remove dead cells, and the number of viable cells was determined by a colorimetric method. This assay is based on the in situ reduction of 3-(4,5-dimethylthizol-2-yl)-2,5-diphenyltetrazolium bromide (MTT) by viable cells (36). After addition of MTT to a final concentration of 1 mg/ml, the cells were incubated at 37°C for 2 h. The resulting formazan precipitate was dissolved with isopropanol containing 0.04 N HCl, and the absorbance of the solution was determined spectrophotometrically at 570 nm with the use of a Victor3 1420 multilabel counter. The antiviral effect for each virus was expressed as the EC50 (the drug concentration at which 50% of infected cells were protected against CPE).
Virus yield reduction assay. MDCK cells were infected (MOI = 0.001) and after 1 h of virus adsorption, the cells were washed with DPBS and then cultured in DMEM supplemented with 1 μg/ml TPCK-trypsin with or without zanamivir. The final concentrations of the inhibitor were 0, 0.01, 1, and 100 μM. To determine the virus yields, the supernatants from the infected cells were harvested 48 h postinfection, and titers were determined for infectivity in MDCK cells.
RESULTS
First, we generated a series of recombinant viruses that contained the HA gene from the CH/95 virus and the remaining genes from the WSN virus and differed from each other only by the number and location of the potential glycosylation sites near the receptor-binding site (Fig. 1; Table 1). The recombinant viruses were subjected to SDS-PAGE analysis, which revealed the anticipated difference in electrophoretic mobility of the HA0 molecules, depending on the number of available sites for glycan attachment (Fig. 2A). Pretreatment of the samples with an enzyme that removes N-glycans (PNG-ase F) was accompanied by a loss of the difference in the HA0 electrophoretic mobilities (Fig. 2B). Therefore, the results of the SDS-PAGE analysis confirmed that a mutation(s) introduced into the HAs of the recombinant viruses abolished glycan attachment at single, double, or triple sites (94a, 129, and 163).
To determine how the introduced changes in the HA affected the efficiency of binding to cellular receptors by the recombinant viruses, we used the host-specific hemagglutination assay (21). The CH/95 virus, as well as other contemporary human influenza A viruses (34), efficiently agglutinates guinea pig erythrocytes but not chicken erythrocytes. The virus preparations were standardized by HA content with the use of anti-HA antibodies by enzyme-linked immunosorbent assay and tested with guinea pig erythrocytes. The difference in the reciprocal HA titers was 2 fold among the virus preparations (128 to 256 hemagglutination units per 50 μl), which is in a range of this assay error. In contrast, there was a substantial difference in the HA titers determined with the use of chicken erythrocytes. The recombinant variant, whose glycans at residues 94a, 129, and 163 had been removed (Table 1; G0) efficiently agglutinated chicken erythrocytes, even at elevated temperatures (37°C). In the presence of glycan at 129, the reciprocal HA titer was reduced by 8 fold. The HA titers were even lower, when glycan at residue 163 was present alone or in combination with the other glycans. In contrast, glycan at residue 94a either did not interfere with heamagglutination (Table 1, G1) or even improved the virus ability to agglutinate chicken erythrocytes (Table 1, G1,2 versus G2).
Assessment of virus susceptibility to NA inhibitors. Because the recombinant viruses shared the NA gene (WSN), we anticipated that they would be equally susceptible to oseltamivir carboxylate by NA inhibition assay. Indeed, the estimated IC50s for the recombinant viruses (n = 8) felt into a narrow range from 0.8 nM to 2 nM (Table 1). However, the susceptibility of the recombinant viruses to NA inhibitor differed dramatically in MDCK cell culture. Based on the results of an MTT assay, the virus lacking all three glycans was highly susceptible to oseltamivir carboxylate (Table 1, G0). Peculiarly, addition of glycan at residue 94a (G1) enhanced by >4 fold the drug susceptibility of the virus in MDCK cells (0.04 μM versus <0.01 μM). In contrast, the presence of glycan at 129 was accompanied by substantial (>100-fold) reduction in the susceptibility to oseltamivir carboxylate, yet this effect was annulled by glycan at 94a (Table 1, G2 and G1,2). The viruses carrying glycan at 163, alone or in a combination with the other two glycans, were resistant to oseltamivir carboxylate (>10,000 fold). Importantly, the MTT assays conducted in the presence of the other NA inhibitor zanamivir produced similar results (Table 1).
The viruses carrying glycan at 163 were essentially zanamivir resistant, based on the results of the infectious yield reduction assay (Table 1). They exhibited a <10-fold reduction in the virus titers in the cell culture supernatants at the highest drug concentration tested (100 μM). In contrast, the viruses lacking all three glycans or carrying glycan at 94a, exhibited a >5-magnitude reduction, whereas the viruses carrying glycan at 129 exhibited an intermediate drug susceptibility (Table 1). Taken together, these results indicate that glycans attached at the HA receptor-binding site can produce the opposite effects on the virus agglutinating ability and susceptibility to NA inhibitors in MDCK cells.
Next, we assessed whether the HA amino acid backbone would influence the glycans' effect on the virus susceptibility to NA inhibitors. To this end, the HA of the WSN virus, which already contains glycan at 129, was modified to create an additional glycosylation site at residue 94a and/or 163 (Fig. 2C). The IC50s for the generated viruses assessed against oseltamivir carboxylate ranged from 2.0 nM to 3.0 nM by NA inhibition assay (Table 2). By MTT assay, the virus carrying glycans at residues 94a and 129 exhibited an approximately 10-fold increase (Table 2; EC50 = 0.2 μM) in susceptibility to zanamivir and oseltamivir carboxylate compare to the wild type virus, whereas addition of glycan at 163 resulted in drug resistance (EC50, >100 μM). The results of the yield reduction assay produced similar results (Table 2). Specifically, the virus that acquired glycan at 94a was the most susceptible to zanamivir with 10,000-fold reduction in the virus titer (Table 2, G1,2). In contrast, the virus with glycan at 163 showed the least drug susceptibility, with 4-fold reduction in the viral titers at the highest drug concentration tested (Table 2, G2,3). The negative effect of glycan at 163 on drug susceptibility was to some extent neutralized by glycan at 94a (50-fold reduction).
Therefore, we established that the presence of glycan at residue 163 caused drastic reduction of the viruses' susceptibility to NA inhibitors regardless of the HA amino acid backbone (CH/95 or WSN).
Rescue of NA-lacking virus variants. HA glycan at 163 apparently provided a compensatory mechanism for diminished NA activity in the presence of NA inhibitor. The question remained of whether this mechanism would be sufficient to permit NA independence of the virus in cell culture. To test this possibility, reverse genetics experiments were conducted, in which we attempted a rescue of virus variants that were completely devoid of NA activity due to deletion in the NA gene (32). The defective NA gene was modified to encode eGFP instead of the NA ectodomain, which allows monitoring of virus spread in live cell culture (32). At 72 h after transfection, the supernatants were harvested and used to inoculate fresh MDCK cell monolayers. The NA-lacking viruses containing glycan at residue 163 in the HA (CH/95 or WSN) were successfully rescued (Table 3). In contrast, we were unable to rescue NA-lacking viruses whose HAs were missing this glycan.
Effect of HA glycan composition on the virus requirement for NA activity. If NeuAc moieties in the vicinity of the HA receptor-binding site indeed provide a compensation of the lack of NA activity, a carbohydrate structure of glycan at 163 should be critical for release of NA-lacking viruses. Glycoproteins are initially glycosylated by transfer of Glc3Man9GlcNAc2 from a lipid donor to Asn of the protein. The oligosaccharide is subsequently processed to the mature forms containing five to nine mannose residues (high-mannosidic type). Further trimming and elongation reactions in the Golgi apparatus result in formation of complex-type oligosaccharides. Trimming reactions proceed through the action of endomannosidase and Golgi mannosidases I and II. Related and depending on trimming, elongation by the families of N-acetylglucosaminyl-, fucosyl-, galactosyl-, and sialyltransferases takes place (43). dMj, an inhibitor of mannosidase I, blocks formation of complex glycans, resulting in the accumulation of high mannose oligosaccharides with a predominant species being Man9GlcNAc2 (8). Swainsonine, an inhibitor of mannosidase II, also blocks formation of complex glycans and causes accumulation of hybrid-type glycans (17, 30).
MDCK cells were infected with the recombinant G1,2,3 variant and, immediately after virus adsorption, either dMj or swainsonine was added to the medium. At 4 h postinfection, virus proteins were metabolically labeled with [35S]methionine-cysteine in the presence of either dMJ or swainsonine in the medium. The proteins incorporated into the virus particles were analyzed by SDS-PAGE (Fig. 3). The slowest-migrating band was produced by the HA0 from the cells maintained in the absence of N-glycosylation inhibitors (Fig. 3, lane 1), whereas the fastest-moving band was produced by the HA0 synthesized in the presence of dMj (Fig. 3, lane 4). Hydrolysis with endo H produced a substantial shift in the bands synthesized in the presence of inhibitors of N-glycosylation (Fig. 3, lanes 6 and 9), whereas it had a much lesser effect on the electrophoretic mobility of the HA0 synthesized in the absence of the inhibitors (Fig. 3, lane 3). Hydrolysis with PNGase F resulted in a single intense band of the same size that was detected in all three immunoprecipitates.
Having confirmed that inhibitors of N-glycosylation (dMj and swainsonine) arrest the HA glycosylation at different steps, we next tested their effects on the outcome of the rescue experiments. When administered to cell culture medium shortly after transfection, neither inhibitor significantly affected the rescue and the infectious yields of the generated virus with the wild-type NA (Table 4). Arrest of the HA glycosylation at a hybrid step by swainsonine also had no apparent effect on the rescue and yields of the NA activity-lacking virus carrying the wild-type HA and caused a 4- to 25-fold reduction in the quantity of the NA activity-lacking virus carrying a single glycan at 163 (Table 4). In contrast, dMj, which arrests the HA glycosylation at the high-mannosidic step, caused a drastic reduction in the yields of the NA activity-lacking viruses, especially of the virus with the single glycan at 163 (50 TCID50/ml).
DISCUSSION
A balance between the antagonistic HA and NA functions can be altered by various events, such as virus transmission to a new host, reassortment, or therapeutic intervention (22, 33, 49). It has been reported that both HA and NA determine the sensitivity of naturally occurring avian influenza viruses to zanamivir in vitro (1, 2). The results of the present study present direct evidence for the role of the HA glycosylation in the human influenza virus susceptibility to NA inhibitors and thus provide a further insight into the functional balance of HA and NA. Because the yields of the NA-lacking mutants in the supernatants were attenuated by 10 to 100 fold, compared to those of the viruses' wild-type NA, HA glycosylation provided only a partial compensation for the loss of the NA activity. Both the oligosaccharide composition and attachment site of HA glycans affected the virus requirement for NA activity in cell culture. Glycosylation at residue 163 on the tip of the HA molecule provided a compensatory mechanism for the loss of NA activity which appears to stem from the reduced efficiency of the HA binding to cellular receptors (2, 38). In MDCK cells, desialidation of the HA by viral NA occurs on the cell surface (26). Therefore, when virus is grown in the presence of NA inhibitor, only complex and hybrid-type HA glycans would retain the terminal NeuAc moieties, whereas the structure of high-mannose-type glycans should not be affected (7). The HA1 subunit of the contemporary CH/95 strain contains eight potential sites of glycosylation at Asn 21, 33, 63, 94a, 129, 163, 271, and 289 (H3 numbering) (GenBank accession no. AF398878) (52). Some of those glycans appear to be essential for HA functioning (48), and the corresponding glycosylation motifs are conserved in the majority of strains. The A/WSN/33 virus underwent intensive adaptation to mice and various cell cultures; its HA1 glycosylation pattern differs from that of the other viruses: four potential sites of glycosylation are present at Asn 21, 65, 129, and 271 (18). Information on glycan composition for the human influenza virus HA is sparse (31, 37), yet the available data correlate well with our findings. Glycan at 94a is mostly likely to be of a high-mannosidic type (15), whereas glycan attached at 129, at least for WSN virus grown in MDBK cells, belongs to a complex type (31). We demonstrated that glycans attached at 129 and 163 reduced the agglutinating activity toward chicken erythrocytes, with glycan at 163 playing a dominant role. Importantly, when propagated in the presence of NA inhibitor, the virus carrying the glycan 163 and wild-type NA failed to agglutinate chicken erythrocytes at 4°C (results not shown), indicating the role of terminal neuraminic acid moieties. This result is in accordance with the results of the hemadsorption studies on the cell surface-expressed avian HA (38). To abolish a glycosylation motif, Asn 163 was replaced by Gln in the present study. Since this residue is not a part of the receptor-binding site (27), we do not believe that amino acid substitution by itself was responsible for the observed difference in the virus properties. Glycan at 94a, when present alone or in combination with glycans 129 and 163, enhanced virus susceptibility to NA inhibitors (Tables 1 and 2). The molecular mechanism for this differential effect is not clear. For example, glycan at Asn 94a could enhance HA binding efficiency via auxiliary interaction with a high-mannose specific lectin expressed on the cell surface (i.e., VIP36 of MDCK cells) (14).
We addressed the question of importance of oligosaccharide chain composition in compensating for the loss of the NA activity in the reverse genetics rescue experiments, when rescue of the NA activity-lacking viruses was attempted in the presence of N-glycosylation inhibitors. Swainsonine, which arrests N-glycosylation at the hybrid step, did not substantially affect the rescue of the NA activity-lacking viruses. In contrast, either we were unable to rescue NA-lacking virus or its titer was significantly reduced in the presence of dMj (Table 4). This could be explained by the ability of hybrid type oligosaccharides to retain the terminal NeuAc acid moieties. The other, less-likely explanation could be the difference in a size between a hybrid and a high-mannosidic oligosaccharide chain. We used dMj at a concentration of 15 μg/ml, which efficiently arrested N-glycosylation at a high-mannosidic step (Fig. 3); however, it may be insufficient to completely prevent formation of complex glycans (9). It is unlikely that the negative effect of dMj on the rescue of the NA-lacking mutant was due to general toxicity or reduction of viral infectivity, since no titer reduction was observed for the virus with wild-type NA in the present study (Table 4) or even when MDCK cells were cultured at much greater dMj concentrations (1 mM) (D.F. Smee, personal communication). Acquisition of glycans at 129 and 163 at the receptor-binding site (as well as those at residues 131 and 158 in some viruses) is believed to be a result of antigenic drift of the A/H1N1 viruses in the human population. Masking antigenic epitopes by glycans is one of the mechanisms by which virus escapes neutralization by preexisting antibodies (45). It is believed that a size of glycan is not critical for epitope masking (35). In contrast, glycans containing terminal NeuAc moieties should contribute the most to the virus's ability to escape the antiviral effect of NA inhibitors. Prior studies have found that the recombinant WSN virus carrying the HA and NA of the 1918 virus exhibited oseltamivir susceptibility in MDCK cells (47). Of note, this virus (42) had a glycosylation site at 94a but lacked sites at position 129 and 163. Such viruses therefore should be more susceptible to NA inhibitors than the modern human viruses of the H1N1 subtype, at least in MDCK cells.
In our study, glycan at 163 produced a compensatory effect, regardless of the HA amino acid sequences (WSN or CH/95). Because the HA sequences of these two viruses differed substantially due to antigenic drift, our data indicated that glycans attached at the tip of the HA molecule have the potential to affect susceptibility of a broad spectrum of influenza viruses, even beyond that of the H1 antigenic subtype. Our findings warrant a more detailed analysis of the HA glycosylation effect on human virus susceptibility to NA inhibitors.
ACKNOWLEDGMENTS
This work was supported by U.S. Public Heath Service research grant AI45782 from the National Institute of Allergy and Infectious Diseases.
We are grateful to E. Hoffmann and R. G. Webster, St. Jude Children's Research Hospital, for providing a set of eight plasmids encoding the genome of A/WSN/33 (H1N1) virus, which we used in the reverse genetic experiments.
REFERENCES
Baigent, S. J., R. C. Bethell, and J. W. McCauley. 1999. Genetic analysis reveals that both haemagglutinin and neuraminidase determine the sensitivity of naturally occurring avian influenza viruses to zanamivir in vitro. Virology 263:323-338.
Baigent, S. J., and J. W. McCauley. 2001. Glycosylation of haemagglutinin and stalk-length of neuraminidase combine to regulate the growth of avian influenza viruses in tissue culture. Virus Res. 79:177-185.
Basak, S., M. Tomana, and R. W. Compans. 1985. Sialic acid is incorporated into influenza hemagglutinin glycoproteins in the absence of viral neuraminidase. Virus Res. 2:61-68.
Baum, L. G., and J. C. Paulson. 1991. The N2 neuraminidase of human influenza virus has acquired a substrate specificity complementary to the hemagglutinin receptor specificity. Virology 180:10-15.
Chu, V. C., and G. R. Whittaker. 2004. Influenza virus entry and infection require host cell N-linked glycoprotein. Proc. Natl. Acad. Sci. USA 101:18153-18158.
Colman, P. M. 1989. Neuraminidase: enzyme and antigen, p. 175-210. In R. M. Krug (ed.), The influenza viruses. Plenum Press, New York, N.Y.
Elbein, A. D. 1991. Glycosidase inhibitors: inhibitors of N-linked oligosaccharide processing. FASEB J. 5:3055-3063.
Elbein, A. D., G. Legler, A. Tlusty, W. McDowell, and R. Schwarz. 1984. The effect of deoxymannojirimycin on the processing of the influenza viral glycoproteins. Arch. Biochem. Biophys. 235:579-588.
Elbein, A. D., J. E. Tropea, M. Mitchell, and G. P. Kaushal. 1990. Kifunensine, a potent inhibitor of the glycoprotein processing mannosidase I. J. Biol. Chem. 265:15599-15605.
Gimsa, U., I. Grotzinger, and J. Gimsa. 1996. Two evolutionary strategies of influenza viruses to escape host non-specific inhibitors: alteration of hemagglutinin or neuraminidase specificity. Virus Res. 42:127-135.
Gubareva, L. V., R. Bethell, G. J. Hart, K. G. Murti, C. R. Penn, and R. G. Webster. 1996. Characterization of mutants of influenza A virus selected with the neuraminidase inhibitor 4-guanidino-Neu5Ac2en. J. Virol. 70:1818-1827.
Gubareva, L. V., M. S. Nedyalkova, D. V. Novikov, K. G. Murti, E. Hoffmann, and F. G. Hayden. 2002. A release-competent influenza A virus mutant lacking the coding capacity for the neuraminidase active site. J. Gen. Virol. 83:2683-2692.
Gubareva, L. V., R. G. Webster, and F. G. Hayden. 2002. Detection of influenza virus resistance to neuraminidase inhibitors by an enzyme inhibition assay. Antiviral Res. 53:47-61.
Hara-Kuge, S., T. Ohkura, H. Ideo, O. Shimada, S. Atsumi, and K. Yamashita. 2002. Involvement of VIP36 in intracellular transport and secretion of glycoproteins in polarized Madin-Darby canine kidney (MDCK) cells. J. Biol. Chem. 277:16332-16339.
Hartley, C. A., D. C. Jackson, and E. M. Anders. 1992. Two distinct serum mannose-binding lectins function as beta inhibitors of influenza virus: identification of bovine serum beta inhibitor as conglutinin. J. Virol. 66:4358-4363.
Hierholzer, J. C., and R. A. Killington. 1996. Virus isolation and quantitation, p. 25-44. In B. W. J. Mahy and H. O. Kangro (ed.), Virology methods manual. Academic Press, London, United Kingdom.
Ho, J. J., R. S. Jaituni, S. C. Crawley, S. C. Yang, J. R. Gum, and Y. S. Kim. 2003. N-Glycosylation is required for the surface localization of MUC17 mucin. Int. J. Oncol. 23:585-592.
Hoffmann, E., G. Neumann, Y. Kawaoka, G. Hobom, and R. G. Webster. 2000. A DNA transfection system for generation of influenza A virus from eight plasmids. Proc. Natl. Acad. Sci. USA 97:6108-6113.
Hoffmann, E., and R. G. Webster. 2000. Unidirectional RNA polymerase I-polymerase II transcription system for the generation of influenza A virus from eight plasmids. J. Gen. Virol. 81:2843-2847.
Inkster, M. D., V. S. Hinshaw, and I. T. Schulze. 1993. The hemagglutinins of duck and human H1 influenza viruses differ in sequence conservation and in glycosylation. J. Virol. 67:7436-7443.
Ito, T., Y. Suzuki, L. Mitnaul, A. Vines, H. Kida, and Y. Kawaoka. 1997. Receptor specificity of influenza A viruses correlates with the agglutination of erythrocytes from different animal species. Virology 227:493-499.
Kaverin, N. V., A. S. Gambaryan, N. V. Bovin, I. A. Rudneva, A. A. Shilov, O. M. Khodova, N. L. Varich, B. V. Sinitsin, N. V. Makarova, and E. A. Kropotkina. 1998. Postreassortment changes in influenza A virus hemagglutinin restoring HA-NA functional match. Virology 244:315-321.
Kaverin, N. V., M. N. Matrosovich, A. S. Gambaryan, I. A. Rudneva, A. A. Shilov, N. L. Varich, N. V. Makarova, E. A. Kropotkina, and B. V. Sinitsin. 2000. Intergenic HA-NA interactions in influenza A virus: postreassortment substitutions of charged amino acid in the hemagglutinin of different subtypes. Virus Res. 66:123-129.
Kim, C. U., X. Chen, and D. B. Mendel. 1999. Neuraminidase inhibitors as anti-influenza virus agents. Antivir. Chem. Chemother. 10:141-154.
Liu, C., M. C. Eichelberger, R. W. Compans, and G. M. Air. 1995. Influenza type A virus neuraminidase does not play a role in viral entry, replication, assembly, or budding. J. Virol. 69:1099-1106.
Luo, C., E. Nobusawa, and K. Nakajima. 2002. Analysis of the desialidation process of the haemagglutinin protein of influenza B virus: the host-dependent desialidation step. J. Gen. Virol. 83:1729-1734.
Matrosovich, M., A. Tuzikov, N. Bovin, A. Gambaryan, A. Klimov, M. R. Castrucci, I. Donatelli, and Y. Kawaoka. 2000. Early alterations of the receptor-binding properties of H1, H2, and H3 avian influenza virus hemagglutinins after their introduction into mammals. J. Virol. 74:8502-8512.
McKimm-Breschkin, J. L. 2000. Resistance of influenza viruses to neuraminidase inhibitors. Antiviral Res. 47:1-17.
McKimm-Breschkin, J. L., T. J. Blick, A. Sahasrabudhe, T. Tiong, D. Marshall, G. J. Hart, R. C. Bethell, and C. R. Penn. 1996. Generation and characterization of variants of NWS/G70c influenza virus after in vitro passage in 4-amino-Neu5Ac2en and 4-guanidino-Neu5Ac2en. Antimicrob. Agents Chemother. 40:40-46.
Merkle, R. K., A. D. Elbein, and A. Heifetz. 1985. The effect of swainsonine and castanospermine on the sulfation of the oligosaccharide chains of N-linked glycoproteins. J. Biol. Chem. 260:1083-1089.
Mir-Shekari, S. Y., D. A. Ashford, D. J. Harvey, R. A. Dwek, and I. T. Schulze. 1997. The glycosylation of the influenza A virus hemagglutinin by mammalian cells. A site-specific study. J. Biol. Chem. 272:4027-4036.
Mishin, V. P., M. S. Nedyalkova, F. G. Hayden, and L. V. Gubareva. 2005. Protection afforded by intranasal immunization with the neuraminidase-lacking mutant of influenza A virus in a ferret model. Vaccine 23:2922-2927.
Mitnaul, L., M. N. Matrosovich, M. R. Castrucci, A. B. Tuzikov, N. V. Bovin, D. Kobasa, and Y. Kawaoka. 2000. Balances hemagglutinin and neuraminidase activities are critical for efficient replication of influenza A virus. J. Virol. 74:6015-6020.
Morishita, T., E. Nobusawa, K. Nakajima, and S. Nakajima. 1996. Studies on the molecular basis for loss of the ability of recent influenza A (H1N1) virus strains to agglutinate chicken erythrocytes. J. Gen. Virol. 77:2499-2506.
Munk, K., E. Pritzer, E. Kretzschmar, B. Gutte, W. Garten, and H. D. Klenk. 1992. Carbohydrate masking of an antigenic epitope of influenza virus haemagglutinin independent of oligosaccharide size. Glycobiology 2:233-240.
Nagai, T., Y. Miyaichi, T. Tomimori, Y. Suzuki, and H. Yamada. 1992. In vivo anti-influenza virus activity of plant flavonoids possessing inhibitory activity for influenza virus sialidase. Antiviral Res. 19:207-217.
Nakamura, K., A. S. Brown, and R. W. Compans. 1980. Glycosylation sites of influenza viral glycoproteins. Tryptic glycopeptides from the A/WSN (H0N1) hemagglutinin glycoprotein. Virology 107:208-221.
Ohuchi, M., R. Ohuchi, A. Fieldmann, and H. D. Klenk. 1997. Regulation of receptor binding affinity of influenza virus hemagglutinin by its carbohydrate moiety. J. Virol. 71:8377-8384.
Palese, P., and R. W. Compans. 1976. Inhibition of influenza virus replication in tissue culture by 2-deoxy-2,3-dehydro-N-trifluoro-acetyl-neuraminic acid (FANA): mechanism of action. J. Gen. Virol. 33:159-163.
Pan, Y. T., H. Hori, and A. D. Elbein. 1987. The effect of glycoprotein-processing inhibitors on the secretion of glycoproteins by Madin-Darby canine kidney cells. Biochem. Cell Biol. 65:345-353.
Penn, C. R., L. M. Barnett, R. C. Bethell, R. Fenton, K. L. Gearing, N. Healy, and A. J. Jowett. 1996. Selection of influenza virus with reduced sensitivity in vitro to the neuraminidase inhibitor GG167 (4-guanidino-Neu5Ac2en): changes in the haemagglutinin may compensate for loss of neuraminidase activity, p. 735-740. In L. E. Brown, A. W. Hampson, and R. G. Webster (ed.), Options for the control of influenza III. Elsevier Science B.V., Amsterdam, The Netherlands.
Potier, M., L. Mameli, M. Belisle, L. Dallaire, and S. B. Malencon. 1979. Fluorometric assay of neuraminidase with a sodium (4-methylumbelliferyl-alpha-D-N-acetylneuraminate) substrate. Anal. Biochem. 94:287-296.
Reid, A. H., T. G. Fanning, J. V. Hultin, and J. K. Taubenberger. 1999. Origin and evolution of the 1918 "Spanish" influenza virus hemagglutinin gene. Proc. Natl. Acad. Sci. USA 96:1651-1666.
Roth, J. 2002. Protein N-glycosylation along the secretory pathway: relationship to organelle topography and function, protein quality control, and cell interactions. Chem. Rev. 102:285-303.
Shinya, K., Y. Fujii, H. Ito, T. Ito, and Y. Kawaoka. 2004. Characterization of a neuraminidase-deficient influenza a virus as a potential gene delivery vector and a live vaccine. J. Virol. 78:3083-3088.
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.
Stray, S. J., R. D. Cummings, and G. M. Air. 2000. Influenza virus infection of desialylated cells. Glycobiology 10:649-658.
Tumpey, T. M., A. Garcia-Sastre, A. Mikulasova, J. K. Taubenberger, D. E. Swayne, P. Palese, and C. F. Basler. 2002. Existing antivirals are effective against influenza viruses with genes from the 1918 pandemic virus. Proc. Natl. Acad. Sci. USA 99:13849-13854.
Wagner, R., D. Heuer, T. Wolff, A. Herwig, and H. D. Klenk. 2002. N-Glycans attached to the stem domain of haemagglutinin efficiently regulate influenza A virus replication. J. Gen. Virol. 83:3-9.
Wagner, R., M. Matrosovich, and H. D. Klenk. 2002. Functional balance between haemagglutinin and neuraminidase in influenza virus infections. Rev. Med. Virol. 12:159-166.
Wagner, R., T. Wolff, A. Herwig, S. Pleschka, and H. D. Klenk. 2000. Interdependence of hemagglutinin glycosylation and neuraminidase as regulators of influenza virus growth: a study by reverse genetics. J. Virol. 74:6316-6323.
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.
Winter, G., S. Fields, and G. G. Brownlee. 1981. Nucleotide sequence of the haemagglutinin gene of a human influenza virus H1 subtype. Nature 292:72-75.(Vasiliy P. Mishin, Dmitri)