The binding properties of the H5N1 influenza virus neuraminidase as inferred from molecular modeling

The avian influenza H5N1 virus has emerged as an important pathogen, causing severe disease in humans and posing a pandemic threat. Substrate specificity is crucial for the virus to obtain the ability to spread from avian to human. Therefore, an investigation of the binding properties of ligands at the molecular level is important for understanding the catalytic mechanism of the avian influenza virus neuraminidase and for designing novel and specific inhibitors of H5N1 neuraminidase. Based on the available crystal structure of H5N1, we have characterized the binding properties between sialic acid, methyl 3’sialyllactoside, methyl 6’sialyllactoside and the H5N1 influenza virus neuraminidase using molecular docking and molecular dynamics simulations. Obtained molecular dynamics trajectories were analyzed in terms of ligand conformations, N1-ligand interactions, and in terms of loop flexibility. It was found that in the N1-SA complex the sialic acid ring undergoes a transition from the B _2,5 to the ² C ₅ conformation. However, in the N1-3SL and N1-6SL complexes sialic acid remained in the distorted boat conformation. The obtained results indicate that 3SL has only weak interactions with the 150-loop, whereas the N1-6SL complex shows strong interactions. Most of the differences arise from the various conformations around the glycosidic linkage, between the sialic acid and galactose, which facilitate the above interactions of 6SL with the enzyme, and as a consequence the interactions between the 150- and 430- loops. This finding suggests that the altered flexibility of loops in and around the active site is one of the reasons why the avian N1 preferentially cleaves sialic acid from α-(2-3)-Gal glycoconjugates over α-(2-6)-Gal. These molecular modeling results are consistent with available experimental results on the specificity of N1.     

Theoretical Investigation on the Binding Specificity of Sialyldisaccharides with Hemagglutinins of Influenza A Virus by Molecular Dynamics Simulations

Recognition of cell-surface sialyldisaccharides by influenza A hemagglutinin (HA) triggers the infection process of influenza. The changes in glycosidic torsional linkage and the receptor conformations may alter the binding specificity of HAs to the sialylglycans. In this study, 10-ns molecular dynamics simulations were carried out to examine the structural and dynamic behavior of the HAs bound with sialyldisaccharides Neu5Acα(2–3)Gal (N23G) and Neu5Acα(2–6)Gal (N26G). The analysis of the glycosidic torsional angles and the pair interaction energy between the receptor and the interacting residues of the binding site reveal that N23G has two binding modes for H1 and H5 and a single binding mode for H3 and H9. For N26G, H1 and H3 has two binding modes, and H5 and H9 has a single binding mode. The direct and water-mediated hydrogen bonding interactions between the receptors and HAs play dominant roles in the structural stabilization of the complexes. It is concluded from pair interaction energy and Molecular Mechanic-Poisson-Boltzmann Surface Area calculations that N26G is a better receptor for H1 when compared with N23G. N23G is a better receptor for H5 when compared with N26G. However, H3 and H9 can recognize N23G and N26G in equal binding specificity due to the marginal energy difference (≈2.5 kcal/mol). The order of binding specificity of N23G is H3 > H5 > H9 > H1 and N26G is H1 > H3 > H5 > H9, respectively. The proposed conformational models will be helpful in designing inhibitors for influenza virus.     

Alterations in receptor binding properties of recent human influenza H3N2 viruses are associated with reduced natural killer cell lysis of infected cells

Natural killer (NK) cell recognition of influenza virus-infected cells involves hemagglutinin (HA) binding to sialic acid (SA) on activating NK receptors. SA also acts as a receptor for the binding of influenza virus to its target host cells. The SA binding properties of H3N2 influenza viruses have been observed to change during circulation in humans: recent isolates are unable to agglutinate chicken red blood cells and show reduced affinity for synthetic glycopolymers representing SA-alpha-2,3-lactose (3'SL-PAA) and SA-alpha-2,6-N-acetyl lactosamine (6'SLN-PAA) carbohydrates. Here, NK lysis of cells infected with human H3N2 influenza viruses isolated between 1969 and 2003 was analyzed. Cells infected with recent isolates (1999 to 2003) were found to be lysed less effectively than cells infected with older isolates (1969 to 1996). This change occurred concurrently with the acquisition of two new potential glycosylation site motifs in HA. Deletion of the potential glycosylation site motif at 133 to 135 in HA1 from a recent isolate partially restored the agglutination phenotype to a recombinant virus, indicating that the HA-SA interaction is inhibited by the glycosylation modification. Deletion of either of the recently acquired potential glycosylation sites from HA led to increased NK lysis of cells infected with recombinant viruses carrying modified HA. These results indicate that alterations in HA glycosylation may affect NK cell recognition of influenza virus-infected cells in addition to virus binding to host cells.     

Two glycosylation sites in H5N1 influenza virus hemagglutinin that affect binding preference by computer-based analysis

Increasing numbers of H5N1 influenza viruses (IVs) are responsible for human deaths, especially in North Africa and Southeast Asian. The binding of hemagglutinin (HA) on the viral surface to host sialic acid (SA) receptors is a requisite step in the infection process. Phylogenetic analysis reveals that H5N1 viruses can be divided into 10 clades based on their HA sequences, with most human IVs centered from clade 1 and clade 2.1 to clade 2.3. Protein sequence alignment in various clades indicates the high conservation in the receptor-binding domains (RBDs) is essential for binding with the SA receptor. Two glycosylation sites, 158N and 169N, also participate in receptor recognition. In the present work, we attempted to construct a serial H5N1 HA models including diverse glycosylated HAs to simulate the binding process with various SA receptors in silico. As the SA-α-2,3-Gal and SA-α-2,6-Gal receptor adopted two distinctive topologies, straight and fishhook-like, respectively, the presence of N-glycans at 158N would decrease the affinity of HA for all of the receptors, particularly SA-α-2,6-Gal analogs. The steric clashes of the huge glycans shown at another glycosylation site, 169N, located on an adjacent HA monomer, would be more effective in preventing the binding of SA-α-2,3-Gal analogs.     

Cleavage mechanism of the H5N1 hemagglutinin by trypsin and furin

The cleavage property of hemagglutinin (HA) by different proteases was the prime determinant for influenza A virus pathogenicity. In order to understand the cleavage mechanism, molecular modeling tools were utilized to study the coupled model systems of the proteases, i.e., trypsin and furin and peptides of the cleavage sites specific to H5N1 and H1 HAs, which constitute models of HA precursor in complex with cleavage proteases. The peptide segments 'RERRRKKR downward arrow G' and 'SIQSR downward arrow G' from the high pathogenic H5N1 H5 and the low pathogenic H1N1 H1 cleavage sites were docking to the trypsin and furin active pockets, respectively. It was observed through the docking studies that trypsin was able to recognize and cleave both the high pathogenic and low pathogenic hemagglutinin, while furin could only cleave the high pathogenic hemagglutinin. An analysis of binding energies indicated that furin got most of its selectivity due to the interactions with P(1), P(4), and P(6), while having less interaction with P(2) and little interactions with P(3), P(5), P(7), and P(8). Some mutations of H5N1 H5 cleavage sequence fitted less well into furin and would reduce high pathogenicity of the virus. These findings hint that we should focus at the subsites P(1), P(4), and P(6) for developing drugs against H5N1 viruses.     

Structural modeling of the distamycin A-d(CGCGAATTCGCG)2 complex using 2D NMR and molecular mechanics

The structure of the distamycin A-d(CGCGAATTCGCG)2 complex has been determined through a combination of SKEWSKY and NOESY 2D NMR experiments and molecular mechanics calculations. NMR data provided upper bounds on many proton-proton pairs. The advantage of the SKEWSKY/NOESY method is that small groups of strongly coupled spins can be treated accurately as isolated systems. The AMBER molecular mechanics package, modified to include the NMR constraints, was used in energy refinements. Distamycin A fits snugly into the 5'-AATT-3' minor-groove binding site. Structural analysis revealed van der Waals contacts between A5, A6, and A18 C2H and drug H3 protons, potential three-center hydrogen bonding between drug amide protons and adenine N3 and thymine O2 atoms analogous to the spine of hydration in the crystal structure of the free DNA, and stacking of the sugar O1' atoms of A6-C21, T7-T20, and, T8-T19, over drug pyrrole rings 1, 2, and 3, respectively. In addition to hydrophobic effects, hydrogen bonding, and electrostatic interactions proposed by others, it is suggested that stacking interactions between DNA sugar O1' atoms and the three drug pyrrole rings contribute to the stability of the complex.     

Identification of amino acid residues of influenza A virus H3 HA contributing to the recognition of molecular species of sialic acid

To identify a determinant of human H3 hemagglutinin (HA) amino acid residues linked to the recognition of molecular species of sialic acid, we generated six mutant viruses possessing either the wild-type HA gene from A/Memphis/1/71 (H3N2) or a genetically single-mutated HA gene at position 137, 144, 155, 158 or 193 from a genetic backbone of A/WSN/33 (H1N1) by reverse genetics. We evaluated the binding ability with four types of synthetic sialylglycolipids. The results indicate that the amino acid substitutions Thr155 to Tyr and Glu158 to Gly in H3 HA facilitate virus binding to N-glycolylneuraminic acid.     

Evaluations for In Vitro Correlates of Immunogenicity of Inactivated Influenza A H5, H7 and H9 Vaccines in Humans

Background

Serum antibody responses in humans to inactivated influenza A (H5N1), (H9N2) and A (H7) vaccines have been varied but frequently low, particularly for subunit vaccines without adjuvant despite hemagglutinin (HA) concentrations expected to induce good responses.

Design

To help understand the low responses to subunit vaccines, we evaluated influenza A (H5N1), (H9N2), (H7N7) vaccines and 2009 pandemic (H1N1) vaccines for antigen uptake, processing and presentation by dendritic cells to T cells, conformation of vaccine HA in antibody binding assays and gel analyses, HA titers with different red blood cells, and vaccine morphology in electron micrographs (EM).

Results

Antigen uptake, processing and presentation of H5, H7, H9 and H1 vaccine preparations evaluated in humans appeared normal. No differences were detected in antibody interactions with vaccine and matched virus; although H7 trimer was not detected in western blots, no abnormalities in the conformation of the HA antigens were identified. The lowest HA titers for the vaccines were <1∶4 for the H7 vaccine and 1∶661 for an H9 vaccine; these vaccines induced the fewest antibody responses. A (H1N1) vaccines were the most immunogenic in humans; intact virus and virus pieces were prominent in EM. A good immunogenic A (H9N2) vaccine contained primarily particles of viral membrane with external HA and NA. A (H5N1) vaccines intermediate in immunogenicity were mostly indistinct structural units with stellates; the least immunogenic A (H7N7) vaccine contained mostly small 5 to 20 nm structures.

Summary

Antigen uptake, processing and presentation to human T cells and conformation of the HA appeared normal for each inactivated influenza A vaccine. Low HA titer was associated with low immunogenicity and presence of particles or split virus pieces was associated with higher immunogenicity.     

An Induced Pocket for the Binding of Potent Fusion Inhibitor CL-385319 with H5N1 Influenza Virus Hemagglutinin

The influenza glycoprotein hemagglutinin (HA) plays crucial roles in the early stage of virus infection, including receptor binding and membrane fusion. Therefore, HA is a potential target for developing anti-influenza drugs. Recently, we characterized a novel inhibitor of highly pathogenic H5N1 influenza virus, CL-385319, which specifically inhibits HA-mediated viral entry. Studies presented here identified the critical binding residues for CL-385319, which clustered in the stem region of the HA trimer by site-directed mutagenesis. Extensive computational simulations, including molecular docking, molecular dynamics simulations, molecular mechanics generalized Born surface area (MM_GBSA) calculations, charge density and Laplacian calculations, have been carried out to uncover the detailed molecular mechanism that underlies the binding of CL-385319 to H5N1 influenza virus HA. It was found that the recognition and binding of CL-385319 to HA proceeds by a process of "induced fit" whereby the binding pocket is formed during their interaction. Occupation of this pocket by CL-385319 stabilizes the neutral pH structure of hemagglutinin, thus inhibiting the conformational rearrangements required for membrane fusion. This "induced fit" pocket may be a target for structure-based design of more potent influenza fusion inhibitors.     

2017–2019年我国南方地区高致病性H5N6亚型禽流感病毒血凝素蛋白分子特征分析

【背景】自2014年以来,H5N6禽流感病毒在我国家禽和活禽市场持续进化,成为人类和动物健康的重大威胁。【目的】对2017–2019年中国南方地区93株高致病性H5N6禽流感病毒的HA基因进行分子进化分析。【方法】接种9–11日龄鸡胚分离核酸检测阳性的H5N6标本,运用下一代测序平台对病毒分离物进行全基因组测序,从NCBI和GISAID数据库下载参考序列,利用BLAST、MEGA6.1及Clustal X等软件进行序列分析。【结果】2017–2019年,从189份江苏省H5亚型禽类/环境标本和1名H5N6患者咽拭子标本中共分离到43株病毒,完成了33株H5N6病毒的全基因组测序。下载网上同时期中国其他地区流行的H5N6毒株序列,对总计93株H5N6病毒的HA基因进行分子进化分析。93株H5N6病毒中有78株属于Clade 2.3.4.4h,9株病毒属于Clade 2.3.4.4e,4株H5N6病毒属于Clade 2.3.4.4b,1株属于Clade 2.3.4.4f,1株属于Clade 2.3.4.4g。所有93株病毒HA蛋白的裂解位点含有多个碱性氨基酸,表明它们都属于高致病性禽流感病...     

Synthesis, biological evaluation, and structural studies on N1 and C5 substituted cycloalkyl analogues of the pyrazole class of CB1 and CB2 ligands

A series of N1 and C5 substituted cycloalkyl and C5 4-methylphenyl analogues of the N-(piperidin-1-yl)-4-methyl-1H-pyrazole-3-carboxamide class of cannabinoid ligands were synthesized. The analogues were evaluated for CB1 and CB2 receptor binding affinities and receptor subtype selectivity. The effects of pyrazole substitution on ligand conformation and as such receptor affinities was not readily apparent; therefore, the geometries of the N1 and C5 substituents relative to the pyrazole ring were studied using high field NMR spectroscopy and systematic molecular mechanics geometry searches. An analysis of the relative ring geometries and functional group orientations provides new insight into the structural requirements of the CB1 and CB2 ligand binding pocket.     

Molecular structure, conformation and interactions of antitumor antibiotic cyanonaphthridinomycin, a covalent binder of DNA

X-ray, NMR and molecular modeling studies on cyanonaphthridinomycin (C22H26N4O5), a DNA binding antibiotic, have been carried out to study the structure, conformation and interactions with DNA. The crystals belong to the space group P21 with the cell dimensions of a = 5.934(1)b = 20.684(4), c = 16.866(3)A, gamma = 90.9 degrees and Z = 4(two molecules/asymmetric unit). The structure was solved by direct methods and difference Fourier methods and refined to an R value of 0.087 for 4061 reflections. The conformation of the molecule is compared with that of naphthridinomycin. There are differences in the orientation of the methoxyl group and the saturated oxazole ring. 1 and 2D NMR studies have been carried out and the dihedral angles obtained from coupling constants have been compared with those obtained from the crystal structure. Molecular mechanics studies were carried out to obtain the energy minimized structure and its comparison with X-ray and NMR results. Molecular modelling studies were performed to propose models for drug-DNA interactions. Both partial intercalation and groove-binding models have been proposed.     

A computational H5N1 neuraminidase model and its binding to commercial drugs

In order to understand the mechanisms of ligand binding and interaction between two commercial drugs (ligands), zanamivir and oseltamivir and H5N1 Influenza Virus Neuraminidase subtype N1, a three-dimensional model of N1-ligand (GenBank accession no. AAS654617) was initially generated by homology modeling using the 13 high-resolution X-ray structures of neuraminidase N2 and N9 as the template. With the aid of the molecular mechanics and molecular dynamics methods, the final implicit solvent refined model was obtained. It was, then, assessed by PROCHECK, PROSA and VERIFY3D. With this model, a flexible docking study was performed. The results show strong hydrogen bond interactions between the glycerol side chains of zanamivir and Arg29 of the N1. Common hydrogen bonds between the carboxyl groups and Arg279 were found for both drugs. It was also found that the Glu30, Asp62, Arg63, Arg204, Trp310, Tyr313, Glu336, Ile338, Trp348, Ala349 were observed to facilitate the enzyme-ligand non-bonding interactions as they are located within the radius of 5 Å from all atoms of both drugs. Charge distribution was evaluated using the semi-empirical AM1 method. The results show that the total net charges of the –NH side chain of zanamivir is less negative than that of oseltamivir. This is in contrast to what is observed for the amide and alkyl (ether/glycerol) side chains. In comparison of the binding free energies between the X-ray N2-ligand and N9-ligand complexes, N1-ligand binding is found to be less potent than N2 and N9 subtypes, while N2-ligand and N9-ligand are roughly comparable. In addition, it is interesting to observe that the binding free energies for all three subtypes of the zanamivir complexes are lower than those of oseltamivir.     

H9N2 virus-derived M1 protein promotes H5N6 virus release in mammalian cells: Mechanism of avian influenza virus inter-species infection in humans

H5N6 highly pathogenic avian influenza virus (HPAIV) clade 2.3.4.4 not only exhibits unprecedented intercontinental spread in poultry, but can also cause serious infection in humans, posing a public health threat. Phylogenetic analyses show that 40% (8/20) of H5N6 viruses that infected humans carried H9N2 virus-derived internal genes. However, the precise contribution of H9N2 virus-derived internal genes to H5N6 virus infection in humans is unclear. Here, we report on the functional contribution of the H9N2 virus-derived matrix protein 1 (M1) to enhanced H5N6 virus replication capacity in mammalian cells. Unlike H5N1 virus-derived M1 protein, H9N2 virus-derived M1 protein showed high binding affinity for H5N6 hemagglutinin (HA) protein and increased viral progeny particle release in different mammalian cell lines. Human host factor, G protein subunit beta 1 (GNB1), exhibited strong binding to H9N2 virus-derived M1 protein to facilitate M1 transport to budding sites at the cell membrane. GNB1 knockdown inhibited the interaction between H9N2 virus-derived M1 and HA protein, and reduced influenza virus-like particles (VLPs) release. Our findings indicate that H9N2 virus-derived M1 protein promotes avian H5N6 influenza virus release from mammalian, in particular human cells, which could be a major viral factor for H5N6 virus cross-species infection.     

Conformation of the branched O-specific polysaccharide of Shigella dysenteriae type 2: molecular mechanics calculations show a compact helical structure exposing an epitope which potentially mimics galabiose

Conformational analyses of the branched repeating unit of the O-antigenic polysaccharide of Shigella dysenteriae type 2 have been performed with molecular mechanics MM3. A filtered systematic search on the trisaccharide alpha-D-GalNAc-(1-->3)-[alpha-D-GlcNAc-(1-->4)]-alpha-D-GalNAc forming the branch, shows essentially a single favored conformation. Also, the downstream alpha-D-GalNAc-(1-->4)-alpha-D-Glc linkage is sterically constrained. The alpha-D-Glc-(1-->4)-beta-D-Gal moiety, however, forms a more flexible link region between the branch points, and shows a 90 degrees bend similar to what is known for the galabiose moiety occurring in globo-glycolipids. The calculations indicate that consecutive repeating units in their minimum energy conformation arrange in a helical structure with three repeating units per turn. This helix is very compact and appears to be stabilized by hydrophobic interactions involving the N-acetyl groups at the branch points. Random conformational search suggests the existence of another helical structure with four repeating units per turn. It appears possible that the alpha-D-Glc-(1-->4)-beta-D-Gal moiety, which is exposed on the surface of the helical structures, can evade recognition by the immune system of the host by the mimicry of globo structures.     

Homology modeling and examination of the effect of the D92E mutation on the H5N1 nonstructural protein NS1 effector domain

Virulent H5N1 strains of influenza virus often harbor a D92E point mutation in the nonstructural protein NS1. This crucial mutation has been correlated with increased virulence and/or cytokine resistance, but the structural implications of such a change are still unclear. Furthermore, NS1 protein could also be a potential target for the development of novel antiviral agents against H5N1 strains. Therefore, a reasonable 3D model of H5N1 NS1 is important for the understanding of the molecular basis of increased virulence and the design of novel antiviral agents. Based on the crystal structure of a non-H5N1 NS1 protein, a model of H5N1 NS1 was developed by homology modeling, molecular mechanics and molecular dynamics simulations. It was found that the D92E mutation could result in weakened interactions of the carboxylate side chain with other phosphorylated residues, thereby activating phosphorylation of NS1. Figure Superposition of snapshots picked from the two molecular dynamic (MD) trajectories: a H5N1 NS1 homology model and b non-H5N1 NS1 crystal structure after 0 (green ribbon), 5 (blue ribbon) and 10 ns (pink ribbon) MD simulation     

Conformation, hydrogen bonding and aggregate formation of guanosine 5'-monophosphate and guanosine in dimethylsulfoxide.

The tetrabutylammonium salt of guanosine 5'-monophosphate (5'-GMP) dissolves in DMSO-d6 forming aggregated species which exhibit some properties of reverse micelles. 1H NOESY experiments show that the 5'-GMP adopts the syn conformation about the glycosidic bond. Molecular mechanics calculations reveal a stable structure with this conformation in which the phosphate group and the amino group of the base are in close enough proximity to hydrogen bond. In contrast inosine 5'-monophosphate in DMSO-d6, which has no NH2 group for hydrogen bond stabilization of the syn conformation, is shown by NMR to have the anti structure. Guanosine in DMSO-d6 behaves differently from 5'-GMP. Guanosine adopts the anti conformation and forms a symmetric dimer via hydrogen bonding between the N3 and NH2 of the bases.     

Site-directed fragment-based generation of virtual sialic acid databases against influenza A hemagglutinin

In this study fragment-based drug design is combined with molecular docking simulation technique, to design databases of virtual sialic acid (SA) analogues with new substitutions at C2, C5 and C6 positions of SA scaffold. Using spaces occupied by C2, C5 and C6 natural moieties of SA when bound to hemagglutinin (HA) crystallographic structure, new fragments that are commercially available were docked independently in all the pockets. The oriented fragments were then connected to the SA scaffold with or without incorporation of linker molecules. The completed analogues were docked to the whole SA binding site to estimate their binding conformations and affinities, generating three databases of HA-bound SA analogues. Selected new analogues showed higher estimated affinities than the natural SA when tested against H3N2, H5N1 and H1N1 subtypes of influenza A. An improvement in the binding energies indicates that fragment-based drug design when combined with molecular docking simulation is capable to produce virtual analogues that can become lead compound candidates for anti-flu drug discovery program.     

Docking of sialic acid analogues against influenza A hemagglutinin: a correlational study between experimentally measured and computationally estimated affinities

In this study fragment-based drug design is combined with molecular docking simulation technique, to design databases of virtual sialic acid (SA) analogues with new substitutions at C2, C5 and C6 positions of SA scaffold. Using spaces occupied by C2, C5 and C6 natural moieties of SA when bound to hemagglutinin (HA) crystallographic structure, new fragments that are commercially available were docked independently in all the pockets. The oriented fragments were then connected to the SA scaffold with or without incorporation of linker molecules. The completed analogues were docked to the whole SA binding site to estimate their binding conformations and affinities, generating three databases of HA-bound SA analogues. Selected new analogues showed higher estimated affinities than the natural SA when tested against H3N2, H5N1 and H1N1 subtypes of influenza A. An improvement in the binding energies indicates that fragment-based drug design when combined with molecular docking simulation is capable to produce virtual analogues that can become lead compound candidates for anti-flu drug discovery program.     

N-Acetylneuraminyllactosylceramide, GM3-NeuAc, a new influenza A virus receptor which mediates the adsorption-fusion process of viral infection. Binding specificity of influenza virus A/Aichi/2/68 (H3N2) to membrane-associated GM3 with different molecular species of sialic acid

Agglutinates of native chicken erythrocytes caused by influenza virus A/Aichi/2/68 (H3N2) at 4 degrees C were potently fused and lysed at low pH (optimum pH 5.3) at 37 degrees C. Exogenous gangliosides GM3 (Sia alpha 2-3Gal beta 1-4Glc beta 1-ceramide) and GM2 (GalNAc beta 1-4(Sia alpha 2-3)-Gal beta 1-4Glc beta 1-ceramide) were integrated into the membranes of chicken asialoerythrocytes within 5-min incubation at 37 degrees C. We found that the incorporation of ganglioside GM3 containing N-acetylneuraminic acid into asialoerythrocytes restored the biological responsiveness to the virus as established by agglutination at 4 degrees C and fusion and hemolysis at 37 degrees C at pH 5.3. Biological responsiveness of GM3-NeuAc-erythrocytes to the virus was considerably higher than that of GM3-NeuGc-erythrocytes under the same experimental conditions. Treatment of the GM3-NeuAc-erythrocytes with neuraminidase again resulted in the complete abolishment of the response to the virus. Erythrocytes containing GM2-NeuAc showed no detectable biological responses toward the virus. The above results indicate that the hemagglutinin of influenza virus A/Aichi/2/68 (H3N2) recognizes the sialyloligosaccharide chain of ganglioside GM3 as its receptor which mediates the adsorption and fusion process on the virus entry into the host cells and has more preferential specificity for binding to N-acetylneuraminic acid-containing GM3 than that to N-glycolyl type in the target cell membranes.