Influenza virus neuraminidase with hemagglutinin activity

VIROLOGY

137.314-323 (1934)

Influenza Virus Neuraminidase

with Hemagglutinin

Activity

W. G. LAVER,**’ P. M. COLMAN,t R. G. WEBSTER,* V. S. HINSHAW,$ AND G. M. AIR$ *John Curtin School of Medical Research, Australian National University, Canberra ACT 2601, Australia; jCSIR0, Division of Protein Chemistry, 3@ Royal Parade, Pa&v&, Victoria 3062, Australia; *Department of Virology and Molecular Biology, St. Jude Children’s Research Hospital, P.O. Box 318, Memphis, Tennessee 38101; and ODepartment of Microbiology, University of Alabama, Birmingham, Alabama 3529.4 Received March 26, 1984 accepted May 25, 198.4 Isolated intact influenza virus neuraminidase (NA) molecules of the N9 subtype have been found to possess hemagglutinin (HA) activity which, at equivalent protein concentration, was fourfold higher than that of isolated hemagglutinin molecules of the H3 subtype. The amino-terminal sequence of the N9 NA is the same as in neuraminidases of the eight other influenza A virus NA subtypes previously reported. Viruses possessing N9 NA therefore have two different HA activities and antibody to either HA or NA alone was incapable of inhibiting hemagglutination by the virus. However, antibody to the HA of an HlN9 virus neutralized its infectivity as effectively as it neutralized HlNl or HlN2 viruses whose neuraminidases have no HA activity. (Antibodies to N9 NA did not neutralize the infectivity of viruses with N9 neuraminidase.) 2-deoxy-2,3-dehydroN-acetyl-neuraminic acid inhibited N9 NA activity but had no effect on the HA activity of the isolated N9 NA. One interpretation of this result would be that the HA and NA activities are located in separate sites. Pronase-released N9 NA heads form crystals suitable for X-ray diffraction studies and preliminary data to 2.9 A establish the space group as cubic, 1432 with cell dimension a = 134 A. Data extend to beyond 1.9 A resolution, and these will be collected in the future. INTRODUCTION

The hemagglutinin (HA) and neuraminidase (NA) activities of influenza virus have been shown to be associated with two quite different glycoprotein molecules on the surface of the virus. This was based on findings with human influenza type A viruses of the Hl, H2, and H3 subtypes and the Nl and N2 subtypes (Laver and Kilbourne, 1966) and type B (Lee) virus (No11 et a& 1962; Laver, 1963). On the other hand, in paramyxoviruses the influenza HA and NA activities reside in the same HN glycoprotein molecule (Scheid and Choppin, 19’74a, b; Scheid et al, 1972). The influenza HA monomer is synthesized as a single polypeptide chain which undergoes post-translational proteolysis 1 Author to whom requests for reprints should be addressed. 0042-6322/34 $3.00 Copyright All rights

0 1994 by Academic Press, Inc. of reproduction in any form reserved.

314

to remove an N-terminal signal sequence and to cleave the molecule into two polypeptide chains called HA1 and HA2, with molecular weights of 36,000 and 27,000, respectively. HA1 and HA2 remain joined by a single disulfide bond and each HA “spike” is a trimer of these HA1 and HA2 chains. The HA is responsible for the attachment of virus to cell membrane receptor molecules which contain sialic acid. This interaction varies in detailed specificity for different influenza viruses (Carroll et cd, 1981). The HA also possesses pHdependent fusion properties which facilitate penetration of the virus into the cell during the initial stages of infection (Maeda et ah, 1981). The role of the NA in the infection process is not clear. The enzyme catalyses cleavage of the a-ketosidic linkage between terminal sialic acid and an adjacent sugar residue. This reaction permits

HEMAGGLUTINATING

transport of the virus through mucin and destroys the hemagglutinin receptors on the host cell, thus allowing elution of progeny virus particles from infected cells. (For recent reviews see Colman, 1984 and Colman and Ward, 1934). The removal of sialic acid from the carbohydrate moiety of newly synthesized hemagglutinin and neuraminidase is also necessary to prevent self-aggregation of the virus (Palese et aL, 1974). In general, then, the role of neuraminidase may be to facilitate mobility of the virus both to and from the site of infection, The NA exists as a mushroom-shaped spike on the surface of the virus particle. It has a box-like head made out of four coplanar and roughly spherical subunits and a centrally attached stalk. At the base of the stalk is a sequence of hydrophobic amino acids which are embedded in the virus membrane (Laver and Valentine, 1969; Wrigley et ak, 1973; Wrigley, 1979). Intact NA molecules can be isolated from detergent-disrupted virus particles (Laver and Kilbourne, 1966). NA “heads” can be isolated by pronase digestion of virus (Seto et aL, 1966). NA heads of subtype N2 have been crystallized (Laver, 1978) and their three-dimensional structure has been determined (Varghese et al., 1933). NA molecules of the Nl and N2 subtypes of type A influenza virus and of type B (Lee) virus are not able to agglutinate red cells. However, we have now found that detergent-released neuraminidase molecules from an avian influenza virus of the N9 subtype possess hemagglutinin activity and this paper presents some data on N9 neuraminidase and the biological properties of influenza viruses possessing two different HA activities. MATERIALS

AND

METHODS

VimLses. Type A influenza viruses of the HllN9 subtype were isolated from apparently healthy noddy terns (Anous minutes) on North West Island of Australia’s Great Barrier Reef in December 1975 (Downie et al, 1977). A recombinant virus was prepared having NWS (Hl) hemagglutinin

NEURAMINIDASE

315

and N9 neuraminidase from one of the tern virus isolates (Webster, 1970). This virus, designated NW&-G7OCn (HlN9) was used in the experiments reported in this paper. Isolation of neuraminidase. Intact N9 neuraminidase molecules were isolated from the HlN9 recombinant virus by electrophoresis on cellulose acetate strips after disruption of the virus particles with SDS (Laver and Kilbourne, 1966). Neuraminidase heads were isolated from the HlN9 recombinant virus following pronase digestion of the virus (Laver, 1978). Virus particles were digested with pronase at 20’ for 16 hr and virus cores were removed by centrifugation at 35,000 rpm (Spinco SW 50.1 rotor) for 60 min. The neuraminidase heads in the supernatant were pelleted by centrifugation at 50,000 rpm (Spinco T160 rotor) for 16 hr. The heads were redissolved in saline and further purified by sucrose density gradient centrifugation (Laver, 1978). Fractions off the gradient were dialyzed against distilled water. No crystals were obtained by this procedure (unlike N2 NA), since N9 neuraminidase is completely soluble at low salt concentration. CrystaUization of N9 neuraminidase. The NA was crystallized by mixing equal volumes of NA solution (lo-15 mg/ml in water) and potassium phosphate buffer (1.7 M, pH 6.6). This mixture was equilibrated through the vapor phase with a reservoir of 1.9 M potassium phosphate, pH 6.8. Large rhombic dodecahedral crystals (Fig. 4) with maximum dimensions exceeding 0.5 mm grew in a few days. Amino acid sequence anal+ Aminoterminal sequences of carboxymethylated intact neuraminidase and pronase-released NA heads were obtained using an Applied Biosystems gas-phase sequencer. The PTH-amino acids were identified by HPLC using an Altex Ultrasphere Cl8 15cm column with a methanol gradient (Bhown et aL, 1978), and a Perkin-Elmer LC85B variable-wavelength detector. In our system (Perkin-Elmer series 4 pump), PTH-Val coeluted with diphenylthiourea and PTH-carboxymethylcysteine with phenylisothiocyanate. These amino

316

LAVER

ET

AL.

G70C/75 (HllN9), was isolated from a acids were therefore confirmed by running a second sample with an acetonitrile granoddy tern (Anous minutus) on Australdient over the same column. ia’s Great Barrier Reef. A recombinant (reassortant) virus was prepared having Serological assays. Hemagglutinin (HA) and hemagglutination-inhibition (HI) as- A/NWS hemagglutinin and the tern virus says were done at 0’ as described (Fazekas neuraminidase. Particles of this recomde St Groth and Webster, 1966). Neurbinant virus, NW&-G~OCN (HlN9) were aminidase assays were done as described disrupted with SDS and the neuramini(Aymard-Henry et a& 1973). Enzymedase was isolated by electrophoresis on linked immunosorbent assays (ELISA) cellulose acetate strips (Fig. 1). The NA was eluted from the strips with were done as described (Kida et al, 1982). saline, precipitated with 2 vol of ethanol The assays were carried out in 96-well tissue culture plates (Nunc) coated with at -2O”, redissolved in and dialyzed antigen. Purified viral antigen was dis- against saline. Following the removal of rupted in 0.5 M Tris, pH 7.8, containthe detergent in this way, the NA aggreing 0.5% Triton X-100 and 0.6 M KC1 gated to form rosettes (Fig. 1, insert) and these rosettes were found to possess high and diluted in phosphate-buffered saline (PBS), pH 7.5, to a concentration equivalevels of hemagglutinin activity. lent to 200 HAU of intact virus/50 ~1. The Biological actiwitg per milligram of isodisrupted virus was dispensed at 50 rl/ lated protein. The HA activity of intact well and antigen allowed to adsorb to the N9 NA was similar to that of H3 HA. plastic for 2 hr at 4’. Following adsorption Table 1 shows the HA activities/mg of the coating solutions were removed and protein. N9 NA was isolated from the SDS-disrupted recombinant virus NW&replaced with 100 ~1 of 1% bovine serum albumin (BSA) in PBS. After 1 hr, the G70CN (HlN9) and H3 HA was isolated plates were washed four times with PBS in a similar way from the SDS-disrupted containing 0.05% Tween 20 fourfold serial recombinant virus A/MEM/1/71n-BEL/ dilutions of ascitic fluid in PBS containing 42N (H3Nl) (Laver et ak, 1974). The HA 0.5% BSA and 0.05% Tween 20 (BSAassays were done at 4” in order to prevent of red cell receptors by the PBST) was then added (50 pi/well). Di- destruction N9 NA. At 4’, N9 NA had 2.4 X lo6 HAU/ lutions of ascitic fluid containing antibody to A/Bangkok/l/79 HA were included as mg protein and H3 HA had 8.2 X 10’ a control together with BSA-PBST only. HAU/mg protein, The HA activity of the After overnight incubation at 4”, the N9 neuraminidase was therefore fourfold plates were washed and 50 ~1 of rabbit higher than that of influenza virus hemanti-mouse globulin conjugated to horse- agglutinin of the H3 subtype at 4”. Similar radish peroxidase was added. After 1 hr, assays using N2 NA isolated from the recombinant virus X-7Fl (NW&-RI/5&) the plates were again washed and bound conjugated antibody detected by the ad- (HlN2), showed that this neuraminidase dition of 100 ~1 of substrate solution (0.05 possessed no detectable HA activity (Table M citrate buffer, pH 4.0, 0.008% hydrogen 1). The neuraminidase activities of N9 NA peroxide; 40 mM azino-di-3-ethyl-benzoand N2 NA were similar with fetuin subthiazobine-6-sulfuric acid). The green color strate. that developed was measured by optical N9 NA rosettes adsorbed totally to red density at 405 nm in a multichannel pho- cells at 4” and eluted at 37’, whereas N2 tometer, NA rosettes did not adsorb to a measurThe inhibitor, 2,3-dehydro-2-desoxy-Nable extent. Pronase-released N9 NA acetyl-neuraminic acid was obtained from heads did not have HA activity when Boehring Mannheim, West Germany. mixed with red cells at 4’ and did not adsorb to a measurable extent. Therefore, the HA site on the molecule was destroyed RESULTS by pronase or the site was still intact, but Isolation of intact NA. The virus the binding to cell receptors of individual used in these experiments, A/Tern/Au&/ NA molecules on their own was too weak

HEMAGGLUTINATING

NEURAMINIDASE

317

+VE

NP

ORIGIN

NA

FIG. 1. Isolation of intact N9 neuraminidase molecules. The recombinant virus NWSn-G7OCn (HlN9) was disrupted with SDS and the proteins were separated by electrophoresis on cellulose acetate strips: a strip stained with Coomassie blue is shown. The neuraminidase molecules eluted from unstained strips aggregated by their hydrophobic membrane attachment sequences at the end of the stalks, when the SDS was removed, forming the rosettes seen in the electron micrograph (right). The N9 NA rosettes possessed HA activity: N2 NA rosettes prepared in the same way possessed no detectable HA activity. The specimen was negatively stained with 2% sodium silicotungstate, pH 7. Bar represents 30 nm. Electron micrograph was taken by N. G. Wrigley.

TABLE BIOLOGICAL

1

ACTIVITIES OF ISOLATED AND NA PROTEINS Activity/mg

Isolated protein

Virus source

N9 NA H3 HA N2 NA

NWS/33-G?OC/75 Mem/1/71-Be1142 X-7Fl (HlN2)

HA units” (4Y 2.4 X lo6 8.2 x 106 0

HA

of protein NA units” WO) 6.6 x 10’ 0’ 4.0 x 10’

“Hemagglutination titers (log3 were measured as described (Fazekas de St. Groth and Webster, 1966) at 4’ to prevent destructing red cell receptors by the NA. bNeuraminidase assays were done as described (AymardHenry et aL, 1973). “0 = not detected.

to allow adsorption to occur. If the former explanation is correct, pronase-released N9 NA heads should not compete with N9 NA rosettes for receptors on the red cell, but they should compete if the latter explanation is correct. The HA activity of N9 NA rosettes was therefore measured in the presence of N9 pronase-released NA heads. At a ratio of heads to rosettes of 125:l (based on NA activity), the heads completely blocked the ability of the rosettes to agglutinate red cells, indicating the HA site on the NA heads was still present. Demonstration that the HA activity of isolated G70C (Ns) NA is not due to contaminating HA protein Intact N9 neuraminidase molecules were isolated from

318

LAVER

SDS-disrupted virus particles by electrophoresis on cellulose acetate strips. The neuraminidase was very well separated from the other virus proteins (Fig. 1) and electron micrographs of the isolated NA did not show the presence of any hemagglutinin molecules. Polyacrylamide gel electrophoresis of the isolated intact NA revealed a single polypeptide of about 75,000 mol wt and no trace of protein corresponding to either of the hemagglutinin polypeptides HA1 or HA2 (Fig. 2). Pronase-released N9 NA heads ran as a single band of MW 55,000 consistent with the removal of the stalk region of the NA molecule. Although these tests indicate that the isolated N9 neuraminidase contains no detectable HA protein derived from the recombinant virus used to prepare the NA, it is still possible that the HA activity of the isolated NA could be due to small amounts of contaminating HA protein. However, ELISA tests (Fig. 3) showed that while hyperimmune rabbit antiserum prepared against the isolated N9 neuraminidase rosettes contained high levels of antibody to N9 NA, no antibody which specifically combined with NWS HA could be detected. This finding provides further evidence that the HA activity of the isolated N9 NA is due to the NA itself and not to contaminating HA. Action of antibodies cm the two hemagglutinating activities on the intact virus. Since N9 neuraminidase will agglutinate erythrocytes, any influenza virus carrying this NA will have two hemagglutinating activities. The question therefore arises whether antibodies to one of these HA activities will block the other HA as well. Hemagglutination-inhibition assays (Table 2) indicate that antibodies to the HA are relatively inefficient at stopping the virus hemagglutinating by means of its NA and vice versa. Hyperimmune antiserum to the isolated G70C neuraminidase had an HI titer of l/10,000 to the isolated neuraminidase but a titer of only l/l60 when this neuraminidase was associated with a hemagglutinin in the virus particle. Similarly, monoclonal antibodies to the hemagglutinin of NWS that inhibited the

ET AL. NA

NA

HA

HEADS

FIG. 2. Polyacrylamide gel electrophoresis of intact N9 NA isolated from the SDS-disrupted recombinant virus (Fig. 1); pronase-released N9 HA “heads” and intact H3 hemagglutinin isolated from the SDSdisrupted recombinant virus A/MEMPHIS/l/‘IluBEL/43n (H3Nl). The right-hand lane contains molecular weight markers, top to bottom, phosphorylase B (92,500), BSA (66,200), ovalbumin (45,000), a-chymotrypsinogen (25,700), p-lactoglobulin (l&400), and cytochrome c (12,300).

homologous virus to a titer of 112500 only inhibited a recombinant virus containing NWS hemagglutinin and G70C neuraminidase (HlN9) to a titer of l/320. These experiments show that the presence of a hemagglutinating neuraminidase in an influenza virus isolate can complicate identification of that virus if monospecific antisera to the HA are used in the diagnostic tests. Neutralization of irlfectivity of viruses possessing N9 neuraminidase. Since anti-

HEMAGGLUTINATING 14

0 Isolated G70C neuromimdase (N91 0 NWS/Bongkok iH1N2) o AlMal/NY/6750/78(HZN2)

1.2 F

2

3 4 ANTIBODY

5 6 7 DILUTION (loglo)

B

FIG. 3. ELISA binding curves showing that hyperimmune rabbit antiserum raised against the isolated G70C (N9) NA rosettes contains a high level of antibodies to the NA, but no antibodies specific for NWS (Hl) HA of the recombinant virus from which the NA was isolated. Some antibody bound to the virus with NWS HA: similar levels bound to the heterologous virus (A/MAL/NY/6750/78) with H2 HA and N2 NA and it is assumed that this antibody bound to the carbohydrate component of the HA in each case.

bodies directed to the HA of A/NWS/nG70CN (HlNQ) virus did not effectively inhibit hemagglutination by this virus (because of the presence of the hemagglutinating neuraminidase), we asked whether these antibodies were capable of neutralizing the infectivity of this virus. TABLE

NEURAMINIDASE

319

The results (Table 3) showed that monoclonal antibodies to the HA neutralized viruses possessing N2 and N9 with equal efficiency. On the other hand, antibodies to N2 or N9 NA were not able to neutralize infectivity of these viruses possessing these neuraminidases. Failure to inhibit hemagglutinin activity on N9 neuraminidase with an analog of N-acet&euraminic acid that inhibits enzyme activity. The above results show that intact N9 neuraminidase will agglutinate red cells very efficiently. We do not know, however, whether the HA activity is due to the enzyme active site of the NA binding to its substrate or whether the HA and NA activities reside in different sites on the NA molecule. Early studies suggested that on the HN surface glycoprotein of paramyxoviruses, the hemagglutinin and neuraminidase activities resided at the same site on the molecule (Scheid and Choppin, 1974b). However, Portner (1981) found that the neuraminidase inhibitor, 2-deoxy-2,3-dehydroN-acetyl-neuraminic acid would inhibit enzyme activity of Sendai virus HN glycoprotein without inhibiting its hemagglutinin activity, suggesting that two sites are present. This inhibitor was tested for its ability to inhibit the NA and HA activities of the NQ neuraminidase (Table 4). 2-deoxy2,3-dehydro-N-acetyl-neuraminic acid in-

2

HEMAGGLUTINATION-INHIBITION TISERA WHEN TESTED AGAINST POSSESSING N9 NEURAMINIDASE

TABLE

(HI)

TITERS OF ANINFLUENZA VIRUSES

NEUTRALIZATION OF INFECTIVITY OF AN INFLUENZA VIRUS POSSESSINGN9 NEURAMINIDASE Dilution of antibody (log,,,) neutralizing the following viruses

HI titers with the following antigens G70C

Antibodies to Isolated G70C NA (N9) (rabbit) NWS virus HA (Hl) (hybridoma)

isolated NA N9

NWS-G’IOC (HlN9)

NWS (HlNl) Antibodies

10,000 40

160 320

3

160 2500

Note. The reagents usedin the HI assaywere all equilibrated to 0’ in an ice bath before use and the tests were done in plastic trays at 0’ as described by Fazekas and Webster, 1966.

to

NWS (Hl) (hybridoma) Isolated G70C (N9) (rabbit) N2 (hybridoma)

X-7F, (HlN2)

NWSG70C (HlN9)

4.0

4.5

4.0 1.3

1.5 4.0

Note. The figures give the dilution of antibodies that cause 50% reduction in plaque number on MDCK cells as described (Kida et al, 1980).

LAVER

320

ET AL.

TABLE

4

THEEFFECTOF~-DEOXY-~-~-DEHYDRO-N-ACETYL-NEURAMINICACIDONTHEHEMAGGLUTININAND NEURAMINIDASE ACTIVITIESOFISOLATEDNQ NEURAMINIDASE

Concentration of inhibitor Of)

Inhibition of NA activity” (%) 0 7 17 27 38 62 73 82 75

(iy h310

5.44 -c 5.47 5.47 -

HA (37” 30+ min) log,,

HA (37” 80+ min) log,,

HA (37” llO+ min) log,,




’ NA activity was determined on fetuin substrate. ‘HA activity of the NA was measured in the presence of the inhibitor at 0” and 37”. The red cells + HA + inhibitor mixture were incubated at 37” for the times shown, the cells were then allowed to settle and the HA titers recorded. ‘Not done.

hibited the NA activity of N9 neuraminidase (fetuin substrate) at a concentration of approximately lo-’ M, although at the highest concentration tested, the inhibition of NA activity was only 82%. On the other hand, the HA activity of the NA was not inhibited at all at any inhibitor concentration tested (Table 4) and consequently NA adsorbed to red cells in the presence of inhibitor but did not elute except after prolonged incubation at 37” (Table 4). These results suggest that the HA and NA activities may reside in different sites on the molecule. Alternatively, a single site may be involved and the inhibitor may act by allowing the NA to bind substrate but blocking its catalytic activity. N-terminal amino acid sequence. Nine antigenically distinct neuraminidase subtypes of type A influenza virus have been isolated. For subtypes l-8, the N-terminal amino acid sequences of the NA (deduced from gene sequencing) are totally conserved and read: Met-Asn-Pro-Asn-GlnLys. However, the “universal” primer used to sequence these regions by the dideoxy method failed to give sequence data for

the N9 NA gene (Blok and Air, 1982). The N-terminal amino acid sequence of N9 NA reads Met-Asn-Pro-Asn-Gln-Lys-Ile -Leu-Thr-Ser-Ala-Thr-Ala-Leu-ValIle-Pro-Thr-Ile-Ala-Val-Ile. The blank at position 9 probably represents either cysteine or a glycosylated asparagine residue. The first six residues are the same as in NA subtypes l-8 (Blok and Air, 1982), indicating that in this respect N9 does not differ from the other type A neuraminidases. X-ray d(.fv-action data Pronase-released N9 NA “heads” were crystallized by vapor diffusion from potassium phosphate solution. Large rhombic dodecahedral crystals were obtained (Fig. 4). X-ray diffraction photographs (Fig. 5) establish the space group as cubic, 1432 with cell dimension a = 184 A. The volume of the asymmetric unit of the crystal is 1.3 X 10’ A3 and as such it can accommodate only one 50,000 M, subunit of the neuraminidase tetramer. The fractional volume of the cell occupied by protein is 0.46 and this crystal form is therefore more densely occupied than either of the N2 neuraminidase crystals (Varghese et CAL,1983). Given that the N2 neuraminidase exhibits cir-

HEMAGGLUTINATING

NEURAMINIDASE

321

The N9 NA gene could be readily transferred into a recombinant (reassortment) virus containing NWS (Hl) hemagglutinin and intact N9 NA molecules or pronasereleased NA “heads” were isolated from this HlN9 recombinant by methods developed for neuraminidase of the N2 subtype (Laver and Kilbourne, 1966; Laver, 1978). We believe that the HA activity of the neuraminidase isolated from the SDSdisrupted recombinant virus, NWSHG70h (HlN9), resides in the NA molecule itself and not in contaminating HA for the following reasons:

FIG. 4. Crystals of G7OC (N9) pronase-released NA suitable for X-ray diffraction studies. The crystals were grown from phosphate buffer solution by vapor diffusion. Bar represents 1 mm.

1. Electron micrographs (Fig. 1) showed no HA present in the preparation of NA rosettes which possessed high levels of HA activity. 2. Polyacrylamide gel electrophoresis of the NA rosettes showed no contaminating HA1 or HA2 polypeptides. 3. NA rosettes of the N2 subtypes isolated in the same way as the N9 NA had no demonstrable HA activity. 4. N9 NA rosettes are completely adsorbed by red cells at 4” (and elute quantitatively at 37’) Pronase-released N9 HA heads do not adsorb to red cells at 4”, possibly because the heads do not aggre-

cular fourfold symmetry, the N9 tetramer must be centered on the crystallographic fourfold axes of the 1432 space group. A complete data set to 2.9 A resolution has been collected. Data extend to beyond 1.9 A resolution (Fig. 5) and these will be collected in the future. The tetrameric protein has only two degrees of freedom in the crystal and we expect that an initial solution will be obtained as a result of searching with the present N2 neuraminidase image (Varghese et ah, 1983). DISCUSSION

The experiments reported here show that intact neuraminidase (NA) molecules isolated from a type A influenza virus found in noddy terns on Australia’s Great Barrier Reef possess hemagglutinin (HA) activity. This virus is designated A/Tern/ Au&/G-70C/75 (HllN9).

FIG. 5. X-ray diffraction pattern of G70C NA extends to Bragg spacings of 1.9 A (crystal to film distance 60 mm X = 1.38 A).

322

LAVER

gate, and cooperation between neighboring NA molecules in the rosette may be needed to enable these to attach to red cells. The hemagglutinin has similar properties. Isolated intact HA molecules which form rosettes (Laver and Valentine, 1969) will agglutinate red cells very efficiently, but bromelain-released HA will not hemagglutinate and shows no measurable adsorption to red cells (Laver et a& 1974). Competition tests showed that the N9 NA “heads” were able to inhibit the attachment of the NA rosettes to cells, presumably by competing for receptor sites on the cells. 5. Antisera raised against the N9 NA rosettes did not contain antibodies to the HA of the recombinant virus from which the NA was isolated (these antisera inhibited the NA activity of the isolated NA to high titer). Antisera raised against the isolated Hll HA of the parental HllN9 tern virus also did not inhibit the HA activity of the N9 NA. These experiments show that the HA activity of intact N9 NA resides in the NA molecule, but whether the activity is due to the enzyme active center of the NA (Colman et ab, 1983) combining with its substrate or whether the HA and NA activities reside in separate sites on the NA molecule is not known. Scheid and Choppin (197413) described experiments which suggested that on the HN glycoprotein of paramyxoviruses the HA and NA activities resided in the same site. However, Portner (1981) has obtained evidence which suggests that on the HN protein of Sendai virus (a paramyxovirus), the HA and NA activities reside in separate sites. The NA activity of Sendai virus was inhibited by more than 95% by an analog of N-acetyl neuraminic acid (2deoxy - 2,3 - dehydro - N- acetyl - neuraminic acid) at a concentration of lo-“ M. This analog did not inhibit the HA activity of Sendai virus even at lo-’ M. We have obtained similar results with the isolated N9 neuraminidase (Table 4) and have shown that in the presence of the analog, the N9 NA rosettes will attach to red cells but will not elute at 3’7” except after prolonged incubation. This suggests

ET AL.

that the HA and NA activities of N9 neuraminidase also reside in separate sites on the NA molecule, but the question of whether the HA activity of the NA is due to enzyme-substrate combination or whether two different sites are involved has not been resolved. Heterogeneous antisera to N9 NA inhibit to high titer both HA and NA activities of the NA. Monoclonal antibodies are being prepared to N9 NA and these should resolve the question of whether one site or two different sites are present on the NA. Hemagglutination by viruses possessing N9 neuraminidase was not inhibited by antibodies to the hemagglutinin (because of the presence of the hemagglutinating neuraminidase). Thus, HI tests can give ambiguous results and are unsatisfactory for serological studies with such viruses. Since viruses possessing N9 NA have two hemagglutinating activities, it is conceivable that each will independently attach the virus to host cell receptors. Previous work has shown that only antibodies to the HA will efficiently neutralize virus infectivity and this study (Table 3) also shows that this also applies to a virus possessing a hemagglutinating neuraminidase. The experiments are consistent with the idea that the N9 NA does not play a significant role in the initial stages of virus infection. The N9 NA apparently does not provide an independent receptor binding site for cell attachment and penetration that can function when antibody is attached to HA on the same virus particle. ACKNOWLEDGMENTS This work is supported by Grants AI 08831 and AI 18203 from the National Institute of Allergy and Infectious Diseases, by Childhood Cancer Center Support (CORE) Grant CA 21765, and Comprehensive Cancer Center Grant CA 13148 from the National Cancer Institute, and by ALSAC. This collaborative project was greatly helped by international telephone facilities provided by the Australian Overseas Telecommunications Commission. The authors acknowledge the excellent technical assistance of Mary Ann Bigelow, Ken Cox, and Jean Clark. REFERENCES AYMARD-HENRY, W. R., LAVER,

W.

M., COLEMAN, G., SCHILD,

M. T., DOWDLE, G. C., and WEBSTER,

HEMAGGLUTINATING R. G. (1973). Influenza virus neuraminidase and neuraminidase inhibition test procedures. Bull WHO 48,199-202. BHOWN, A. S., MOLE, J. E., WEISSINGER, A., and BENNETT, J. C. (1978). Methanol solvent for rapid analysis of phenylthiolhydantoin amino acids by high-pressure liquid chromatography. J, Chrometop. 148, 532-535. BLOK, J., and AIR, G. M. (1982). Variation in the membrane-insertion and “stalk” sequences in eight subtypes of influenza type A virus neuraminidase. Biochemistry 21,4001-4007. CARROLL, S. M., HIGA, H. H., and PAULSON, J. C. (1981). Different cell-surface receptor determinants of antigenically similar influenza virus hemagglutinins. J. Biol Chem 266, 8357-8363. COLMAN, P. M. (1984). The structure and function of neuraminidase “Peptide and Protein Reviews,” Vol. 4, in press. COLMAN, P. M., and WARD, C. W. (1984). Structure and diversity of the influenza virus neuraminidase. “Current Topics Microbiology & Immunology,” in press. COLMAN, P. M., VARGHESE, J. N., and LAVER, W. G. (1983). Structure of the catalytic and antigenic sites in influenza virus neuraminidase. Nature (L.imclon) 303, 41-44. DOWNIE, J. C., HINSHAW, V. S., and LAVER, W. G. (1977). The ecology of influenza. Isolation of type ‘A’ influenza virus from Australian pelagic birds. Aust. J. Exp. Biol Med Sci 55, 635-643. FAZEKAS DE ST GROTH, S., and WEBSTER, R. G. (1966). Disquisitions on original antigenic sin. I. Evidence in man. J. Exp Mea! 124,331-345. KIDA, H., BROWN, L. E., and WEBSTER, R. G. (1982). Biological activity of monoclonal antibodies to operationally defined antigenic regions on the hemagglutinin molecule of A/Seal/Massachusetts/ l/80 (H7N7) influenza virus. Virology 122, 38-47. LAVER, W. G. (1963). The structure of influenza viruses. 3. Disruption of the virus particle and separation of neuraminidase activity. Virology 20, 251-262. LAVER, W. G. (1978). Crystallization and peptide maps of neuraminidase “heads” from H2N2 and H3N2 influenza virus strains. virology 86, 78-87. LAVER, W. G., DOWNIE, J. C., and WEBSTER, R. G. (1974). Studies on antigenic variation in influenza virus. Evidence for multiple antigenic determinants on the hemagglutinin subunits of A/Hong Kong/ 68 (H3N2) virus and the A/England/72 strains. Virology 59, 230-244. LAVER, W. G., and KILBOURNE, E. D. (1966). Identification in a recombinant influenza virus of struc-

NEURAMINLDASE

323

tural proteins derived from both parents. Virdogy 30,493~501. LAVER, W. G., and VALENTINE, R. C. (1969). Morphology of the isolated hemagglutinin and neuraminidase subunits of influenza virus. Virology 38, 105-119. MAEDA, T., KAWASAKI, K., and OHNISHI, S. (1981). Interaction of influenza virus hemagglutinin with target membrane lipids is a key step in virusinduced hemolysis and fusion at pH 5.2. Proc. Natl. Acad Sci. USA 78,4133-4137. NOLL, H., AOYAGI, T., and ORLANDO, J. (1962). The structural relationship of sialidase to the influenza virus surface. virology 18, 154-157. PALESE, P., TOBITA, K., UEDA, M., and COMPANS, R. W. (1974). Characterization of temperature sensitive influenza virus mutants defective in neuraminidase. virology 61, 397-410. PORTNER, A. (1981). The HN glycoprotein of Sendai virus: Analysis of site(s) involved in hemagglutinating and neuraminidase activities. Virology 115, 375-384. SCHEID, A., and CHOPPIN,P. W. (1974a). Identification of biological activities of paramyxovirus glycoproteins. Activation of cell fusion, hemolysis and infectivity by proteolytic cleavage of an inactive precursor protein of Sendai virus. Virology 57, 475-490. SCHEID, A., CALIGUIRI, L. A., COMPANS, R. W., and CHOPPIN, P. W. (1972). Isolation of paramyxovirus glycoproteins. Association of both hemagglutinating and neuraminidase activities with the larger SV5 glycoprotein. Virology 50, 640-652. SCHEID, A., and CHOPPIN, P. W. (1974b). The hemagglutinating and neuraminidase protein of a paramyxovirus: interaction with neuraminic acid in affinity chromatography. Virology 62, 125-133. SETO, J. T., DRZENIEK, R., and ROTT, R. (1966). Isolation of a low molecular weight sialidase (neuraminidase) from influenza virus. B&hem. Biophys. Acta 113,402-404. VARGHESE, J. N., LAVER, W. G., and COLMAN, P. M. (1983). Structure of the influenza glycoprotein antigen neuraminidase at 2.9A resolution. Nature (Lmdm) 303, 35-40. WEBSTER, R. G. (1970). Antigenic hybrids of influenza A viruses with surface antigens to order. Virology 42, 633-642. WRIGLEY, N. G. (1979). Electron microscopy of influenza virus. Brit. Mea! Bull 35, 35-38. WRIGLEY, N. G., SKEHEL, J. J., CHARLWOOD, P. A., and BRAND, C. M. (1973). The size and shape of influenza virus neuraminidase. virology 51, 525529.