Properties of an Antigenic Glycoprotein Isolated from Influenza Virus Hemagglutinin

Edward A. Eckert
1973 Journal of Virology  
A purified antigen, HABA protein, has been derived from influenza virus concentrates by extraction with denaturing solvents. The protein lacks hemagglutinating activity but binds completely strain-specific, hemagglutination-inhibiting antibodies and induces neutralizing antibodies in experimental animals. Physicochemical characterization of HABA protein identifies it as a single homogeneous glycoprotein with a molecular weight of 78,000. On dissociation with guanidine or sodium dodecyl sulfate,
more » ... in the presence of reducing agents, only one size of polypeptide with a molecular weight of the order of 40,000 is characteristic of the preparations. The data indicate that HABA protein is a dimer of HA, polypeptide of the influenza virus hemagglutinin substructure, and that only trace amounts of other polypeptides are present. Influenza virus exhibits a complex structure characterized by possession of a lipid protein membrane with spikelike projections extending from the surface. Some seven viral polypeptides have been identified by acrylamide gel electrophoresis, four of which are glycopeptides (4). A major step in establishing the structure of influenza virus would be identification of those polypeptides associated with its several functional subunits. With respect to the substructure involved in hemagglutination, two polypeptides, HA1 and HA2, have been nominated (13). In studies on viral structure, we have isolated a homogeneous protein related to hemagglutinin (7). This protein does not agglutinate erythrocytes but does bind hemagglutination-inhibiting (HI) antibodies. On this .basis it is referred to as hemagglutinin-binding antigen or HABA protein. Evidence has been obtained for the presence of only a single size of polypeptide subunit in HABA protein (9). The present report aims at characterizing the HABA molecule and its constituent polypeptide with reference to composition, molecular dimensions, and immunological properties. The data support the tentative designation of HABA protein as a dimer of HA1 polypeptide arranged in a conformation retaining immunological specificity. MATERIALS AND METHODS Virus. Strain A/PR/8/34 (HON1) was used in most experiments. Virus,was cultivated in embryonated eggs at 35 C, and pteliminary treatment was adsorption elution from erythrocytes. Further purification and concentration was obtained by adsorption elution with BaSO4 (6), followed by centrifugation at 50,000 x g for 60 min. Virus concentrates were prepared by suspending the pellets in 0.89% sodium chloride with 0.08% sodium azide as preservative. To extend the determinations of molecular weights by exclusion chromatography to proteins from other virus strains, A/AA/6/60 (H2N2) and the X7 (Fl) recombinant (HON2) obtained from E. D. Kilbourne (14) were employed in some experiments. For this purpose, virus was pelleted directly from allantoic fluid at 50,000 x g for 60 min and then banded on a 5 to 40% tartrate gradient in an SW25 rotor (Spinco) at 50,000 x g for 60 min. Extraction of HABA protein. A 4-ml amount of virus concentrate, with hemagglutinin (HA) titers of the order of 10 per ml, was mixed with 15 ml of methanol-chloroform (2:1, v/v) and the mixture was shaken gently by hand for 2 min. Next, 5 ml of chloroform was added, and shaking was continued for 2 min. Then, following addition of 5 ml of water and 2 min of shaking, the mixture was centrifuged at 800 x g for 15 min. An insoluble mass, located at the interface, was collected and suspended in 4 ml of water, and a stream of nitrogen was bubbled through the coarse suspension to remove organic solvents. The suspension was centrifuged at 800 x g for 15 min, and the supernatant fluid was discarded. The pellet was 183 on May 9, 2020 by guest http://jvi.asm.org/ Downloaded from ECKERT either lyophilized for future extraction or immediately suspended in 4 ml of solvent containing 7 M guanidine, 0.05 M dithiothreitol (DTT), and 0.1 M tris(hydroxymethyl)aminonrethane-hydrochloride (Tris-hydrochloride) (pH 7.8), and stirred at room temperature for 7 or 16 hr, during which time the suspension became essentially clear. On dialysis in the cold against PBS (0.01 M phosphate buffer [pH 7.4] and 0.1 M NaCl), insoluble proteins flocculated from solution. Dialysis was continued for 48 hr against three changes of PBS, 1 liter each. Preparations were clarified twice by centrifugation at 50,000 x g for 1 hr. The clear supernatant fluid is the HABA protein preparation referred to in this presentation. The sediment was either discarded or, if used for analysis, was washed twice by suspension in 30 ml of PBS followed by centrifugation at 50,000 x g for 1 hr. Reagents. Materials and their suppliers were as follows: guanidine HCl, Ultra Pure, and Tris-hydrochloride, Schwarz/Mann, Orangeburg, N.J.; crystallized bovine plasma albumin, Metrix, Armour Pharmaceutical Co., Chicago, Ill.; dithiothreitol, ovalbumin, 'and bovine gamma globulin, Nutritional Biochemical Corp., Cleveland, Ohio; trypsin and chymotrypsin, Worthington Biochemical Corp.; polyacrylamide gel materials, Bio-Rad Laboratories, Richmond, Calif. Protein determinations. Protein was estimated by the method of Lowry et al. (18) with bovine serum albumin as standard. Carbohydrate determinations. The anthrone method of Roe was employed with controls for charring and tryptophan (22) . Incubation was at 80 C, and absorption was measured over the spectral range from 500 to 640 nm. Amino acid analysis. Samples of protein were hydrolyzed with 5.8 N HCl in vacuo after repeated flushing with nitrogen at 110 + 1 C, and after 24 or 48 hr HCl was removed in vacuo. Amino acid composition was determined with a Spinco model 120 automatic recording analyzer, equipped with an electronic integrator. Molecular exclusion chromatography. Columns had diameters of 2.5 cm and lengths of 45 cm and were equipped with flow adaptors (Pharmacia Fine Chemicals). Eluant was run through 40 cm of gel by upward flow with a Delta micro-metering pump (Watson-Marlow, Falmouth, England) at a rate of 12 ml/hr, and fractions were collected in 4-ml volumes. For determinations in saline, Sephadex G150 and G200 were equilibrated with eluant: 0.1 M KCl, 0.05 M Tris-hydrochloride (pH 7.8), and sodium azide, 0.08%. Void volumes were determined with Blue Dextran 2,000 and internal volumes with 2, 4-dinitrophenol (DNP)-alanine. Protein samples containing 1 to 2 mg were diluted in or dialyzed against eluant. Elution volumes were determined by spectrophotometric measurements of the fractions at 280 nm. For chromatography in the presence of guanidine, large-pore agarose (Bio-Gel A-50m) was equilibrated against eluant consisting of 6 M guanidine, 0.01 M DTT, 0.1 M KCl, 0.05 M Tris-hydrochloride (pH 7.8), and 0.01 M ethylenediaminetetraacetic acid (EDTA). Void volumes were determined with glutaraldehyde-fixed E. coli suspensions and internal volumes with DNP-alanine. Protein samples containing 1 to 2 mg were dissolved in eluant in which the guanidine concentration was increased to 7 M. Effluent fractions were analyzed by ultraviolet (UV) absorption at 280 nm and then dialyzed against PBS, the UV absorption was rechecked, and the blocking antigen titers were measured. Calculation of internal volumes, distribution coefficients, and Stokes radii followed the procedure of Andrews (1). Preparations of HABA protein gave two estimates of elution volumes, one based on blocking antigen titers and the other on UV absorption. The latter values, since they could be estimated with greater precision, were employed to determine distribution coefficients and molecular weights. Polyacrylamide gel electrophoresis. Formulation of the 7.5% acrylamide-0.1% sodium dodecyl sulfate (SDS) gel system followed the methods of Maizel et al. (19) . Samples were dissolved in solvent containing 1% SDS and 1% DTT and were boiled for 2 min. Short-term electrophoresis was in 0.5 by 6 cm gels at 3 mA per gel for 7 hr, and longer runs were in 0.5 by 11 cm gels at 3 mA per gel for 18 hr. Gels were fixed in 50% trichloroacetic acid for 25 hr and then held 8 hr in a solvent containing 45.4% (v/v) methanol and 9.2% acetic acid (w/v) in water. Overnight staining was with 0.25% Coomassie blue in the latter solvent. Gels were destained with two charges of the same solvent, followed by several changes of 7.5% acetic acid in water (w/v). Densitometer tracings were made with the acrylamide gel accessory of a Beckman Acta spectrophotometer. Antisera. Antisera to whole virus were obtained by intravenous inoculation of rabbits with 0.5 ml of virus concentrate containing 4,000 HA units of virus treated with Formalin (1:4,000). Inoculation was repeated after 1 wk, and the rabbits were bled from the ear vein 1 wk later. Booster doses were given at longer intervals, and animals were again bled 1 wk after booster. HA and HI tests. Plastic trays were used with reaction mixtures containing 0.2 ml of hemagglutinin, 0.2 ml of antiserum or PBS, and 0.4 ml of 0.5% chicken red blood cells. Each test was carried out in a conventional manner (21). Blocking antigen test. Stock virus was standardized to 4 HA units per 0.2 ml and antiviral serum to 4 HI units per 0.2 ml. To 0.2 ml of successive twofold dilutions of protein was added 0.2 ml of antiserum. After 30 min at room temperature, 0.2 ml of virus was added to each reaction mixture. After an additional 30 min, during which virus could react with unbound antibody, each cup received 0.2 ml of 1% chicken erythrocytes. Titers were read as the highest dilution of protein giving a positive HA reaction (7). Immunoaffinity columns. Antisera were first chromatographed on diethylaminoethyl (DEAE)cellulose to remove HA inhibitor (17). Anti-PR8 184 J. VIROL. on May 9, 2020 by guest http://jvi.asm.org/ Downloaded from VOL 11, 1973 GLYCOPROTEIN FROM INFLUENZA VIRUS HEMAGGLUTININ serum was dialyzed against 0.017 M phosphate buffer (pH 6.3) and clarified at 800 x g for 10 min. DEAE-cellulose, (DE-52, Microgranular preswollen, Whatman) was equilibrated against the same solvent and poured as a 1.2 by 27 cm column with an overlay of G-25 Sephadex. A 3-ml amount of antiserum, followed by 0.017 M phosphate buffer eluant, was run through the column by gravity, and 4-ml fractions were collected. Concentrated reagent was then added to each chromatographic fraction to give an NaCl concentration of 0.1 bovine serum albumin to 0.5% and sodium azide to 0.08%. HI levels of specific antibody of the fractions were measured with PR8 virus and with A/Japan/305/57 (H,N) heated at 56 C for 30 min as indicator virus for nonspecific inhibitor titers. HI titers for PR8 virus reached 256 at the peak, whereas the corresponding inhibitor titer was less than 2. Preparation of HABA immunoadsorbent was based on the method of Axen (2, 5). Sepharose 4B (Pharmacia Fine Chemical) was first equilibrated with water. To 100 ml of Sepharose gel, suspended in 100 ml of water, was added 10 g of cyanogen bromide dissolved in 100 ml of water with rapid mixing by using a magnetic stirrer. The pH was immediately adjusted with 4 M NaOH and maintained at pH 11.0 to 11.3. When the pH stabilized, the gel was washed on a Buchner funnel over Whatman no. 1 filter paper, with a total of 2 liters of 0.1 M sodium bicarbonate. Gel was suspended in 200 ml of 0.1 M sodium bicarbonate adjusted to pH 9 with sodium hydroxide. A 1-ml amount of HABA protein, containing approximately 1.5 mg of protein, was added, and the mixture was stirred gently overnight at 4 C. The gel was washed exhaustively by successive settling and decanting with a total of 8 liters of PBS containing 0.08% sodium azide. Chromatography was performed in a 1.2 by 10 cm column of HABA-Sepharose gel by gravity feed. A 1-ml amount of DEAE-cellulose-fractionated antiserum followed by PBS eluant was run through the column at a rate of 15 ml/hr, and twelve 4-ml fractions were collected. The column was washed with an additional 30 ml, and then the eluant was changed to 6 M potassium iodide (KI) to elute adsorbed antibody. The fractions were dialyzed overnight against PBS. Both PBS and KI eluant fractions were then titered for HI antibodies. Equilibrium filtration. The procedure of Fazekas de St. Groth (10) was adopted to compare antibody bound by virus and protein preparations. Graded dilutions of virus or protein in 1-ml volumes were mixed with 1 ml of rabbit antivirus serum diluted 1:25 with 1% bovine serum albumin in PBS. After 30 min at room temperature, the mixtures were passed through 0.05-Am pore size membrane filters (Millipore Corp.) by positive pressure. Samples were tested to measure the HI titers of the filtrates, and titers of bound antibody were plotted against dilution of antigen. Parallel, best-fitting lines were drawn for each antigen. The regression line for virus was taken as standard, and the distance between the standard line and the lines for each of the other anti-gens was used to determine the antibody-binding ratios. RESULTS Preparation and properties of HABA protein. The procedure for the preparation of HABA protein described under Materials and Methods leads to a separation of viral proteins based on their relative solubility in PBS after lipid extraction and dissociation with guanidine-DTT. The saline-soluble fraction contains the serologically active HABA protein, whereas a major portion of viral protein is in the form of a flocculent precipitate removable by centrifugation. To determine partitioning of viral constituents, protein and hexose content and selected serological activities were determined for HABA preparations, for insoluble sediments, and for original virus concentrates. Four preparations derived from three different concentrates of strain PR8 influenza virus were analyzed. As shown in Table 1 , 15% of total viral protein was found in HABA protein preparations. While the major portion of protein was recovered in the sediment, the total of the two fractions was approximately 10% less than the estimate for virus concentrate. In explanation, while a substance positive in the Lowry test might have been lost on lipid extraction, it seems more likely that clumps of the coarse sediment adhered to the wall while samples were pipetted. Determinations by the anthrone test gave a value of 0.154 mg of hexose per ml for HABA protein whereas that of the sediment was only 0.139, demonstrating that more than half of the protein-bound hexose was associated with HABA protein (Table 1 ). Since the latter had a much lower protein content, the percentage of hexose per unit of protein averaged 11.7 for TABLE 1. Chemical and serological properties of HABA protein, sediment, and original virus concentratea Deteminaion
doi:10.1128/jvi.11.2.183-192.1973 fatcat:5jm2f7b2lzdnlb46yxjaty5d4e