The N2 neuraminidase of human influenza virus has acquired a substrate specificity complementary to the hemagglutinin receptor specificity

VIROLOGY

180, IO-15 (1991)

The N2 Neuraminidase of Human Influenza Virus Has Acquired a Substrate Specificity Complementary to the Hemagglutinin Receptor Specificity LINDA G. BAUM*s’ AND JAMES C. PAULSON+’ Departments

of *Pathology

and tBiologica/

Chemistry,

UCLA School of Medicine,

Los Angeles,

California

90024

Received June 27, 1990; accepted August 3 1, 1990 A survey of 10 human influenza A viruses of the N2 serotype, isolated between 1957 and 1987, has revealed a drift in neuraminidase linkage specificity. While the earliest N2 strains examined exhibit strict specificity for cleavage of the NeuAca2,SGal sequence, N2 isolates from 1967 to 1968 also show limited activity towards the NeuAca2,6Gal linkage. In strains isolated in 1972 and later, the N2 neuraminidase has approximately equal activity towards both types of linkages. The NeuAca2,6Gal linkage cleaved by the later N2 neuraminidases is the preferred receptor determinant of human H2 and H3 hemagglutinins. Thus, the acquired neuraminidase specificity of the later isolates allows elution of bound virus from erythrocytes derivatized to contain the NeuAcaP,BGal linkage, while earlier isolates, which cleave only the NeuAca2,3Gal sequence, fail to elute from these etythrocytes. These results suggest that the observed drift in N2 neuraminidase specificity in the direction of the oreferred H2 and H3 receptor determinant may facilitate release of progeny virus from host cells. 0 1991 Academic Press, Inc.

INTRODUCTION

proteins have now been shown to exhibit specificity for naturally occurring sialoside linkages (for review, see Drzeniek, 1973; Paulson, 1985). In the case of influenza virus neuraminidases, those examined exhibit strict specificity for hydrolysis of the NeuAca2,3Gal linkage (Drzeniek, 1973; Carroll et al., 1981; Cot-field et al., 1983; Colman and Ward, 1985; Cabezas et a/., 1989). Carroll et al. (1981) showed that the interplay of neuraminidase and hemagglutinin specificities could influence the differential elution from erythrocytes of the influenza subtypes Rl/5+ and Rl/5-, early H2N2 viruses described by Choppin and Tamm (1960). The neuraminidases of both subtypes were found to exhibit high specificity for hydrolysis of the NeuAca2,3Gal linkage, while the hemagglutinins of the Rl/5+ and Rl/5subtypes preferentially bound the NeuAccu2,6Gal and NeuAca2,3Gal linkages, respectively. Thus, the action of the neuraminidase efficiently destroys receptors of the Rl/5- subtype, allowing its elution from erythrocytes. In contrast, the neuraminidase fails to cleave the sialic acid linkages which serve as the preferred receptor determinant of the Rl/5+ virus, accounting for the inability of this virus to elute from erythrocytes. Schulman and Palese (1977) demonstrated that a combination of specific hemagglutinin and neuraminidase genes was necessary for growth in MDCK cells. However, the mechanism by which the preferred neuraminidase-hemagglutinin combinations resulted in increased virulence in MDCK cells has not been elucidated. Indeed, there has been little evidence to date to suggest that a complementary specificity between the

The influenza virus neuraminidase is an exoglycosidase which hydrolyzes sialic acid from an a-ketosidic linkage to an adjacent sugar residue (Gottschalk, 1957). It was originally identified as a receptor-destroying enzyme, because sialic acids are the primary receptor determinant of the viral hemagglutinin which mediates attachment of the virus to the cell surface (Hirst, 1942; Burnet and Stone, 1947). In viva, the neuraminidase is postulated to serve several functions. During virus production, it facilitates release of budding virus from the infected cell by hydrolysis of cell surface sialic acids, and removes sialic acids from viral glycoproteins which would otherwise cause hemagglutinin-mediated aggregation (Palese et al., 1974; Griffin and Compans, 1979; Griffin et al., 1983). In addition, neuraminidase cleavage of sialic acid from respiratory tract mucus is thought to prevent virus entrapment in this substance (Burnet, 1948; Reed, 1969; Tabak eta/., 1982). Early observations of Burnet et al. (1946) and Stone and Ada (1952) described a “receptor gradient,” in which various influenza virus strains were found to differ in their abilities to hydrolyze cell surface receptors recognized by the other strains. The results suggested that the virus strains differed in their neuraminidase and hemagglutinin specificities. Both viral glyco’ To whom reprint requests should be addressed. ’ Present address: Cytel, Inc., 1 1099 Torrey Pines Road, La Jolla, CA 92037. 0042-6822/91

$3.00

Copyright 0 1991 by Academic Press, Inc. All rights of reproduction in any form reserved.

10

NEURAMINIDASE

HAS ACQUIRED

influenza virus hemagglutinins and neuraminidases has selective advantage for the virus. There is increasing evidence, however, from examination of over 60 influenza virus strains from 1947 to 1987 (Hl Nl , H2N2, and H3N2), that humans select for a hemagglutinin receptor specificity exhibiting preferential binding of NeuAca2,6Gal linkages (Rogers and Paulson, 1983; Rogers and D’Souza, 1989; J. C. Paulson, N. Couceiro, L. G. Baum, Y. Kawaoka, and R. G. Webster, unpublished results). In this report, we demonstrate that the human influenza N2 neuraminidase has gradually shifted its specificity from preferential hydrolysis of the NeuAca2,3Gal linkage (1957 H2N2 strains) to hydrolysis of both NeuAcor2,3Gal and NeuAccr2,GGal linkages (1972-l 987 strains). The recently acquired ability of the N2 neuraminidase to cleave the NeuAcol2,6Gal sequence corresponds to the ability of later H3N2 strains to elute from the surface of erythrocytes by destroying cell surface receptors. It is proposed that the gradual drift in specificity of N2 neuraminidases towards the NeuAccu2,GGal linkage results in a selective advantage to the virus, through more efficient destruction of cell surface receptors on the infected cell, allowing release of newly formed virus particles. MATERIALS

AND METHODS

Viruses Seed stocks of A/Rl/5+/57 (Rl5+) and A/RI/%/57 (Rl5-) were the gift of Dr. Purnell Choppin, The Rockefeller University. Seed stocks of A/Japan/305/57 and A/Hong Kong/8/68 were obtained from the American Type Culture Collection. The strains A/Tokyo/3/67, A/ Aichi/2/68, A/Udorn/307/72, A/Memphis/l 02172, Al Victorial3175, A/Texas/l/77 and A/Los Angeles12187 were generously provided by Dr. Robert Webster and Dr. Yoshihiro Kawaoka, St. Jude’s Children’s Research Hospital, Memphis, Tennessee. The clonal isolate of AJMemphis/l02/72, M1/5, and Ml/5 HS8, the receptor variant strain derived from M1/5, have been described (Rogers et al., 1983). All viruses were grown in the allantoic sac of lo-day-old embryonated chicken eggs as previously described (Rogers and Paulson, 1983). Preparation

of derivatized

al -acid glycoprotein

Asialo-crl -acid glycoprotein was resialylated with purified Gal/?l,4GlcNAca2,6 sialyltransferase or Galbl, 3(4)GlcNAca2,3 sialyltransferase and CMP-[14C]NeuAc as previously described (Higa and Paulson, 1985; Corfield et a/., 1983). Sialic acid incorporated was 280 nmol/mg protein for the NeuAcor2,6Gal sub-

SUBSTRATE

11

SPECIFICITY

strate and 300 nmol/mg protein for the NeuAccu2,3Gal substrate (sp act 4700 dpm/nmol NeuAc). Preparation

of derivatized

human erythrocytes

Type A erythrocytes were desialylated by incubation with l/ibrio cholerae neuraminidase and resialylated with Gal/?l,4GlcNAcor2,6 sialyltransferase and CMP[14C]NeuAc, as described earlier (Carroll et a/., 1981; Rogers and Paulson, 1983). Derivatized erythrocytes contained 40-70 nmol NeuAc/ml packed erythrocytes. Neuraminidase

assays

Standard neuraminidase assays for quantitation of viral neuraminidase activity used native al -acid glycoprotein as a substrate. Virus concentrate (5 ~1) was incubated with 0.25 mg bovine serum albumin (BSA) and 0.5 mg 0l1 acid glycoprotein in a final volume of 100 PI of 0.1 M sodium cacodylate, pH 6.5. Released sialic acid in the total sample was quantitated by the thiobarbituric acid assay (Warren, 1959). One unit of neuraminidase activity was defined as the amount of enzyme that caused hydrolysis of 1 pmol NeuAc per minute at 37”. To determine neuraminidase linkage specificity, virus (0.03-0.5 mU neuraminidase activity) was incubated with [14C]NeuAc-labeled al -acid glycoprotein (4-6 nmol NeuAc) and 0.25 mg BSA in 50 ~1 of 0.1 M sodium cacodylate buffer, pH 6.5, for 30 min at 37”. Released [14C]NeuAc was separated from the labeled substrate by chromatography on Sephadex G-50 (0.7 X 13-cm column). The 50-~1 sample was applied to the column and the column developed with 4 X 2-ml additions of 0.1 M NaCI. Four 2-ml fractions were collected, mixed with 2 ml Ecoscint, and counted in a Beckman LS-9000 scintillation counter. Fraction 2 contained intact substrate, and fractions 3 and 4 contained free NeuAc. Percentage of the total NeuAc released was calculated as (cpm free)/(cpm bound + free). Hemagglutination

assays

Hemagglutination titers were determined at room temperature in a microtiter system (Cooke Engineering Co.) as described in Carroll et a/., 1981. Titers were read at 1 hr and hemagglutination units expressed as the reciprocal of the maximum dilution of virus that caused complete agglutination. Virus adsorption

and elution from erythrocytes

Virus was adsorbed to 10% suspensions of native or derivatized erythrocytes in 10 mM sodium diethylbarbiturate, 150 mM NaCI, pH 7.0, 5 mg/ml BBS-BSA as described previously (Carroll et a/., 1981) except vol-

12

BAUM AND PAULSON

umes were scaled up 10 times. The virus-erythrocyte mixture was incubated at 37” and, at the indicated times, duplicate 60-~1 aliquots of cell-free supernatant and resuspended cells were removed and assayed for neuraminidase activity as described (Carroll et a/., 1981). Influenza virus neuraminidase derivatized erythrocytes

treatment

of

Derivatized erythrocytes containing the NeuAca2,6Gal linkage (40 nmol NeuAc/ml packed cells) were incubated with Ml/5 as the source of neuraminidase (100 mu/ml packed cells) as described (Carroll et al., 1981), except that the incubation period was 4 hr. After extensive washing with 0.01 M NaHPO,, 0.15 M NaCI, pH 7.0, the cells were used in hemagglutination assays as described above. RESULTS Two N2 strains of human influenza A have different neuraminidase substrate specificities Previous studies on neuraminidase substrate specificity of the human H2N2 strain A/Rl/5+/57 (Rl5+) revealed preferential hydrolysis ofthe NeuAccu2,3Gal linkage (Carroll eta/., 1981). In contrast, preliminary analysis of the neuraminidase specificity of a clonal isolate of A/Memphis/l 02/72 (Ml/5) demonstrated that this strain had equal activity towards both the NeuAca2,GGal and NeuAca2,3Gal linkages (Pritchett, 1987). To further examine this difference, a direct comparison of the activity of Rl5+ and Ml/5 towards the two types of linkages on glycoprotein substrates was made (Fig. 1). Both viruses were able to release [‘“ClNeuAc from the NeuAccu2,3Gal linkage. However, as predicted from the initial findings, Rl5+ demonstrated a very low rate of hydrolysis of the NeuAca2,6Gal linkage (Fig. 1A), while Ml/5 showed approximately equal rates of hydrolysis for the two substrates (Fig. 1C). Both Rl5+ and Ml/5 exhibit the receptor specificity characteristic of human influenza viruses, preferentially binding to cell surface receptors containing the NeuAca2,6Gal linkage (Rogers and Paulson, 1983). Corresponding receptor variant strains of these two viruses, Rl5- and Ml/5 HS8, with hemagglutinins that preferentially bind NeuAca2,3Gal linkages, were also examined for their neuraminidase specificity. As shown in Figs. 1 B and 1 D, both Rl5- and Ml/5 HS8 exhibit the same neuraminidase specificity as their counterparts of opposite receptor specificity, indicating that neuraminidase specificity is independent of hemagglutinin specificity.

s 2

0 100

Lj

75

z 0

50 25 0

0

lhr 2hr

4hr

6hr

1hr 2hr

4hr

6hr

TIME ( hours) FIG. 1. Differential linkage specificities of early and late N2 neuraminidases. Virus strains were compared for the ability to release [“%]NeuAc from al -acid glycoprotein derivatized to contain either the NeuAca2,3Gal linkage (0) or the NeuAca2,BGal linkage (0). Reaction mixtures containing 0.15 mU of viral neuraminidase activity and labeled substrate (6 nmol [‘QNeuAc) were incubated for the indicated times and percentage NeuAc released was calculated as described under Materials and Methods. (A) R15f (A/Rhode Island/ 5+/57; H2N2), (6) Rl5- (A/Rhode island/5/57; H2N2), (C) Ml/5 (A/ Memphis/102/72; H3N2), (D) Ml/HS8 (receptor variant of A/Memphis/102/72; H3N2).

Neuraminidase specificity influenza A viruses

of human N2

To determine whether the dual neuraminidase specificity seen in Ml/5 was characteristic of recent isolates bearing the N2 neuraminidase, the neuraminidase activity of 10 H2 and H3 strains of human influenza virus was examined. A comparison of the rates of hydrolysis of [14C]NeuAc-labeled al-acid glycoprotein containing either the NeuAca2,3Gal or the NeuAca2,GGal sequence is shown in Fig. 2. All virus strains tested hydrolyzed the NeuAca2,3Gal linkage. However, neuraminidase activity towards the NeuAca2,GGal linkage was seen only in strains isolated in 1967 and later. Neuraminidase linkage specificity drift was most dramatic in viruses isolated in 1972 and later, with all strains demonstrating approximately equal activity towards both types of linkages. The acquisition of specificity for the NeuAca2,6Gal linkage can be seen to occur in at least two steps. Viruses isolated in 19671968 have limited ability to cleave the NeuAccu2,6Gal linkage, while those isolated in 1972 and later have dual substrate specificity. Viral elution from erythrocytes It has long been observed that influenza virus strains differ in their ability to elute from erythrocytes (Stone and Ada, 1952; Choppin and Tamm, 1960; Carroll et

NEURAMINIDASE

HAS ACQUIRED

SUBSTRATE

13

SPECIFICITY

40 q NeuAcoZ,3Gal

Virus strains : 1 2 3 4 5 6 7 8 9 10 Yearof Isolation : ‘57 ‘57 ‘67 ‘68 ‘68 ‘72 ‘72 ‘75 ‘77 ‘87 Subtype : LH2N2J LH3N2Time ( minutes) FIG. 2. Human influenza virus N2 neuraminidases demonstrate a drift in linkage specificity which correlates with year of isolation. Neuraminidase activity was assayed by measuring the release of [‘%]NeuAc from enzymatically derivatized oil -acid glycoprotein, as described in Fig. 1. Samples contained sufficient viral neuraminidase activity (0.03-0.5 mu) to yield 15-30s release from the substrate containing the NeuAcu2,3Gal linkage after a 30.min incubation at 37”. Identical quantities of each viral neuraminidase were then added to an equivalent amount of substrate containing the NeuAca2,GGal linkage, and percentage NeuAc released was determrned. Light bars: release from NeuAca2,3Gal linkage. Dark bars: release from NeuAcaZ,GGal linkage. (1) A/Rl/5+/57, (2) A/Japan/ 305/57, (3)AJTokyo/3/67, (4)AIHong Kong/8/68. (5)A/Aichi/2/68, (6) A/Udorn/307/72, (7) A/Memphis/l 02/72, (8) A/Victoria/3/75, (9) Af Texaslli77, (10) A/Los Angeles/2/87.

al., 1981). This finding has been attributed

to differences in both viral neuraminidase and hemagglutinin specificities. Since all the human viruses examined in Fig. 2 preferentially bind to cell surface receptors containing the NeuAcor2,6Gal linkage (Carroll et al., 1981; Rogers and Paulson, 1983; J. C. Paulson, L. G. Baum, Y. Kawaoka, and R. G. Webster, unpublished results), it was of interest to determine if their differing neuraminidase specificities could influence their ability to elute from erythrocytes. Accordingly, Rl5+ and Ml/5 were examined for their ability to elute from both native human erythrocytes and erythrocytes derivatized to contain the NeuAccr2,6Gal linkage. As shown in Fig. 3, Ml/5 rapidly eluted from the erythrocytes, with 47 and 66% of the total virus present in the supernatants of the native and derivatized cells after 3 hr, respectively. In contrast, Rl5+ exhibited minimal elution from either native or derivatized cells. Thus, while the Rl5+ neuraminidase, with little activity towards the NeuAca2,6Gal linkage, could not inactivate the viral receptors, the dual linkage specificity of the Ml/5 neuraminidase allowed rapid release of cell-associated virus. Hemagglutination

of erythrocytes

treated with Ml /5

Prevention of virus rebinding to cells, as well as release of budding virus, are postulated to be conse-

FIG. 3. Viruses with early and late N2 neuraminidases differ in their ability to elute from native and derivatized erythrocytes. The viruses A/Rhode island/5+/57 (0. 0) and A/Memphis/l02/72 (0, W) were adsorbed to native (A) and derivatized (B) erythrocytes, and aliquots (60 ~1)of a 10% suspension of the cells in BBS-BSA were incubated at 37”. At the indicated times, the samples were centrifuged at 1OOOg for 2 min, and the cell free supernatants were removed. The pelleted cells were resuspended in 50 gl of BBS-BSA and equal volumes of cell-free supernatants and resuspended cells were assayed for viral neuraminidase activity, using [3H]sialylactitol as a substrate (Carroll et a/., 198 1). Eluted virus is expressed as a percentage of the total virus adsorbed to the erythrocytes. In B, erythrocytes were derivatized to contain 40 (0) or 70 (m) nmol NeuAciml packed cells, in the NeuAca2,GGal linkage.

quences of the receptor-destroying activity of the influenza virus neuraminidase. To determine the effect of neuraminidase specificity on the availability of receptor determinants which could mediate virus rebinding, Rl5+ and Ml/5 were used as a source of neuraminidase to treat erythrocytes derivatized to contain the NeuAccu2,6Gal linkage, the preferred receptor determinant. Cells treated with Rl5+ agglutinated irreversibly, and could not be dissociated for further testing, indicating that receptor determinants were not destroyed by the Rl5+ neuraminidase (data not shown). However, cells treated with Ml/5 could be resuspended

TABLE 1 HEMAGGLUTINATIONOF RESIALYATEDEFWTHROCMESAFTERTREATMENT WITH Ml/5 NEURAMINIDASE

Virus

Neuraminidase treatmenta

Rl/5+

-

Ml/5

+ -

Hemagglutination titer 64 0 128 0

+ a Details of neuraminidase and Methods.

treatment

described

under Materials

14

BAUM AND PAULSON

after viral elution. As shown in Table 1, treatment of derivatized erythrocytes with Ml/5 abolished subsequent hemagglutination by both Ml/5 and Rl5+. These results indicate that hydrolysis of the NeuAca2,GGal linkages by the Ml/5 neuraminidase removed receptor determinants for both viruses from the etythrocyte surface.

DISCUSSION Previous studies on the influenza virus N2 neuraminidase showed remarkable specificity for the NeuAccu2,3Gal linkage (Drzeniek, 1973; Carroll et al., 1981; Corfield et a/., 1983). However, the results presented here demonstrate that the N2 neuraminidase has undergone a gradual acquisition of the ability to cleave the NeuAcLu2,GGal linkage. Indeed, recent isolates demonstrate equal activity towards the NeuAca2,3Gal and NeuAcor2,6Gal linkages (Figs. 1 I a. While the amino acid sequences of several human N2 neuraminidases are known (Colman and Ward, 1985), the number of amino acid changes among the antigenic subtypes tested is too great to allow definitive identification of residues involved in recognition of the NeuAca2,BGal sequence. However, testing of additional strains with known neuraminidase sequences, in combination with site-directed mutagenesis, may identify those amino acids responsible for the observed drift in linkage specificity. Although the role of neuraminidase substrate specificity in influenza pathogenicity is not established, the acquisition of the ability to cleave the NeuAca2,GGal sequence and its persistence in later strains raises the possibility that this phenotype affords the virus a selective advantage. The linkage specificity drift of the neuraminidase is in the direction of the preferred receptor determinant of H2 and H3 human influenza strains, the NeuAcol2,6Gal sequence. Since the neuraminidase is thought to allow elution of budding virus from receptor sites on the host cell surface (Palese eta/., 1974; Griffin et al., 1983; Els et a/., 1989), more efficient destruction of hemagglutinin receptor determinants would facilitate this process. This hypothesis is supported by the experiments shown in Fig. 3 and Table 1, demonstrating the rapid inactivation of receptor determinants by the Ml/5 neuraminidase, which can cleave the NeuAca2,6Gal sequence. In the human host, a match of hemagglutinin and neuraminidase specificities for the NeuAca2,6Gal linkage may be relevant to the mechanism of viral pathogenicity. Lectin-binding studies suggest that the surface glycoconjugates of human ciliated tracheal cells, the cells infected by influenza virus (Liu, 1954; Dourmash-

kian and Tyrrell, 1970), contain the NeuAca2,GGal linkage (Baum and Paulson, 1990). Efficient cleavage of the NeuAccu2,GGal sequence at the surface of the infected epithelial cell would result in more rapid dissemination of virus to adjacent, uninfected cells. During the acquisition of specificity for the NeuAcLu2,6Gal linkage, activity towards the NeuAca2,3Gal sequence was retained. Analysis of oligosaccharide structures in human respiratory tract mucin has revealed that these structures contain primarily the NeuAcor2,3Gal linkage (Breg et al., 1987). Thus, the conserved neuraminidase activity against the NeuAca2,3Gal sequence would help the virus to escape mucin entrapment. The influenza virus neuraminidase is thought to mediate a number of functions during the infectious cycle of the virus, and has been shown to be important in determining virulence (Schulman and Palese, 1977; Ogawa and Ueda, 1981). The results presented suggest that the neuraminidase specificity drift towards the preferred hemagglutinin receptor specificity may increase infection efficiency by facilitating virus release from infected cells, and thus provide a selective advantage to more recent N2 strains.

ACKNOWLEDGMENTS The authors thank Dr. Sorge Kelm for helpful discussions and preparation of derivatized erythrocytes. We also thank Dr. Yoshihiro Kawaoka and Dr. Robert Webster for virus seed stocks, and Dr. Peter Colman for his suggestions regarding neuraminidase structure. L.G.B. is a recipient of a Young Investigator Award from the National Foundation for Infectious Diseases.

REFERENCES AIR, G. M., and LAVER,W. G. (1986). The molecular basis of antigenic variation in influenza virus. Adv. Virus Res. 31, 53-l 02. BAUM, L. G., and PAULSON,J. C. (1990). Sialyloligosaccharides of the respiratory epithelium in the selection of human influenza virus receptor specificity. Acta Hisrochem., in press. BREG, J., VAN HALBEEK, H., VLIEGENTHART,1. F. G., LAMBLIN, G., HouVENAL, M. C., and ROUSSEL,P. (1987). Structure bf sialyl-oligosaccharides from bronchial mucus glycoproteins in patients suffering from cystic fibrosis. Eur. /. Biochem. 168, 57-68. BURNET, F. M. (1948). Mucins and mucoids in relation to influenza virus action. IV. Inhibition by purified mucoid of infection and hemagglutination with the virus strain WSE. Aust. J. Exp. Biol. Med. Sci. 26, 381-387. BURNET,F. M., MCCREA, J. F., and STONE, J. D. (1946). Modification of human red cells by virus action: I. The receptor gradient for virus action in human red cells. Brit. /. Exp. Pafhol. 27, 228-236. BURNET, F. M., and STONE, J. D. (1947). The receptor-destroying enzyme of V cholerae. Aust. J. Exp. Biol. Med. 25, 227-233. CABEZAS,J., MILICUA, M., BERNAL, C., VILLAR, E., PEREZ,N., and HANNOUN, C. (1989). Kinetic studies on the sialidase of three influenza B and three influenza A strains. Glycoconjugate f. 6, 2 19-227. CARROLL, S. M., HIGA, H. H., and PAULSON, J. C. (1981). Different

NEURAMINIDASE

HAS ACQUIRED

cell-surface receptor determinants of antigenically similar influenza virus hemagglutinins. /. Biol. Chem. 256, 8357-8363. CHOPPIN, P. W., and TAMM, I. (1960). Studies of two kinds of virus particles which comprise influenza virus A, strains. II. Characterization of stable homogeneous substrains in reactions with specific antibody, mucoprotein inhibitors and erythrocytes. /. Exp. Med. 112,895-920. COLMAN, P. M., and WARD, C. W. (1985). Structure and diversity of influenza virus neuraminidase. Cuff. Top. Microbial. Immunol. 114, 177-255. CORFIELD. A. D., HIGA, H. H., PAULSON,1. C., and SCHAUER, R. (1983). The specificity of viral and bacterial sialidases for a(2-3)- and a(2-6).linked sialic acids in glycoproteins. Biochim. Biophys. Acta 744, 121-126. DOURMASHKIAN,R. R., and TYRRELL,D. A. I. (1970). Attachment of two myxoviruses to ciliated epithelial cells. J. Gen. Viral. 9, 77-88. DRZENIEK. R. (1973). Substrate specificity of neuraminidases. Histothem. 1. 5, 271-290. ELS, M. C., JAVER, W. G., and AIR, G. M. (1989). Sialic acid is cleaved from glycoconjugates at the cell surface when influenza virus neuraminidases are expressed from recombinant vaccinia virus. Viralogy 170, 346-351. GOT~SCHALK. A. (1957). The specific enzyme of influenza virus and Vibrio cholerae. Biochim. Biophys. Acta 23, 645-646. GRIFFIN, J. A., BASAK, S., and COMPANS, R. W. (1983). Effects of hexose starvation and the role of sialic acid in influenza virus release. Virology 125, 324-334. GRIFFIN,J. A., and COMPANS, R. W. (1979). Effect of cytochalasin Bon the maturation of enveloped viruses. J. Exp. Med. 150, 379-391. HIGA, H. H., and PAULSON, J. C. (1985). Sialylation of glycoprotein oligosaccharides with fll-acetyl-, N-glycolyl-, and N-O-diacetylneuraminic acid. /. Biol. Chem. 260, 8838-8848. HIRST, G. K. (1942). Adsorption of influenza hemagglutinins and virus by red blood cells. J. Exp. Med. 76, 195-209. LIU, C. (1954). Studies on influenza infection in ferrets by means of fluorescein labelled antibody. J. Exp. Med. 101, 665-676. OGAWA, T., and UEDA, M. (1981). Genes involved in the virulence of an avian influenza virus Virology 113, 304-313.

SUBSTRATE

SPECIFICITY

15

PALESE,P., TOHITA, K., UEDA, M., and COMPANS, R. W. (1974). Characterization of temperature sensitive influenza virus mutants defective in neuraminidase virology 61, 397-410. PAULSON, J. C. In “The Receptors” (P. M. Conn, Ed.), Vol. II, pp. 131-219. Academic Press, Orlando, 1985. PRITCHETT,T. J. (1987). Doctoral Dissertation. REED, S. (1969). Persistant respiratory virus infection in tracheal organ culture. Brit. /. fxp. Patho/. 50, 378-388. ROGERS,G. N., and D’Souz~, B. L. (1989). Receptor binding properties of human and animal H 1 influenza virus isolates. virology 173, 317-322. ROGERS,G. N., and PAULSON,J. C. (1983). Receptor determinants of human influenza virus isolates: Differences in receptor specificity of the H3 hemagglutinin based on species of origin. Virology 127, 361-373. ROGERS,G. N., PRITCHEIT, T. J.. LANE, J. L., and PAULSON,J. C. (1983). Differential sensitivity of human, avian and equine influenza A viruses to a glycoprotein inhibitor of infection: Selection of receptor specific variants. Virology 131, 394-408. Roar, R. (1979). Molecular basis of infectivity and pathogenicity of myxovirus. Arch. Viral. 59, 285-298. SCHNEIR, M. L., and RAFELSON,M. E. (1966). Isolation and characterization of two structural isomers of N-acetyl-neuraminyllactose from bovine colostrum. Biochim. Biophys. Acfa 130, l-l 1. SCHOLTISSEK,C., ROHDE, W., VON HOYNINGEN,V., and ROTT, R. (1978). On the origin of the human influenza virus subtypes H2N2 and H3N2. \/iro/ogy 87, 13-20. SCHULMAN, 1. L., and PALESE,P. (1977). Virulence factors of influenza A viruses: WSN Virus neuraminidase required for plaque production in MDBK cells. /. Viral. 24, 170-l 76. STONE, J. D., and ADA, G. L. (1952). Electrophoretic studies of virusred cell interaction: Relationship between agglutinability and electrophoretic mobility. Brit J. Exp. Patho/. 33, 428-439. TABAK, L. A., LEVINE, M. J., MAMDEL, I. D., and ELLISON, S. A. (1982). Role of salivary mucins in the protection of the oral cavity. /. Oral Patho/. 11, 1-17. WARREN, L. (1959). The thiobarbituric acid assay of sialic acids. /. Biol. Chem. 234, 1971-1975.