Inactivation of Influenza B Virus by Normal Guinea-pig Serum

J. gen. Virol. (1987), 68, 1135-1141. Printed in Great Britain 1135 Key words: influenza B virus/inhibitor/guinea-pig serum Inactivation of Influenza B Virus by Normal Guinea-pig Serum ByFUMIYAMAMOTO, 1 KOICHIROMAENO, 1*SHINICHIROSHIBATA, ~ MASAO IINUMA, 2 AKIO MIYAMA 3 AND YASUKO KAWAMOTO 3 Laboratory of Virology, Research Institute for Disease Mechanism and Control Nagaoya University School of Medicine, Nagaoya, 2Gamagori Fukashi Hospital Gamagori City and 3Department of Microbiology, Fujita Gakuen University School of Medicine, Toyoake City, Japan (Accepted 22 December 1986) SUMMARY Normal guinea-pig serum (GPS) lacking detectable antiviral antibody efficiently neutralized the infectivity of influenza B virus grown in chick embryos or MDCK cells. The inhibitor was heat-labile and sensitive to trypsin digestion. This fl-like inhibitor required Ca 2+ and the complement components C 1 and C4 for its activity. In contrast, GPS did not inactivate influenza A virus. Influenza B virus from which the neuraminidase activity of the spikes on the viral envelope had been eliminated by trypsin digestion was also inactivated to a level comparable to untreated virus. Complement component C1 alone bound directly to influenza B virus and inhibited its haemagglutinin activity, We suggest that the fl-like inhibitor in GPS is a component of the classical complement pathway which is triggered by the protein moiety of influenza B virus haemagglutinin, leading to virus neutralization. INTRODUCTION A variety of normal animal sera contain non-specific inhibitors of influenza viruses. They have been classified as ~, fl and 7 inhibitors, depending on their chemical composition and properties (Krizanova & Rathova, 1969). The c~ inhibitor is a sialylated glycoprotein which inhibits haemagglutination by influenza A and B viruses but, unlike fl and ~ inhibitors, does not prevent infection (Francis, 1947). Its activity is destroyed by neuraminidase (NA) but not by heating for 30 min at 56 C. The ~ inhibitor was first recognized as a non-specific inhibitor of an H2N2 strain of influenza A virus (Shimojo et al., 1958; Sugiura et al., 1961) and resembles c~ inhibitor in most of its properties. In contrast, the fl inhibitor is apparently not sialylated and is inactivated by heat (56 C for 30 rain) but not by NA (Chu, 1951). It has been reported that certain enveloped viruses interact with the complement system, without the participation of antibody, to activate the classical complement pathway and/or the alternative complement pathway, leading to direct inactivation of the viruses (Cooper, 1979, 1984; Sissons, 1984). Judging by physicochemical and biological similarities it is most likely that the fl inhibitor is part of the complement system. Although influenza A and B viruses resemble each other in many structural and biological properties, the inhibitor present in normal animal sera (such as human and mouse sera) neutralizes influenza A viruses but has little effect on influenza B viruses (Chu, 1951 ; Ananthanaryan & Paniker, 1960; Krizanova & Rathova, 1969). This has led us to investigate whether the fl-like inhibitor against influenza B virus is present in normal animal sera and whether complement is able to inactivate this virus. We present evidence that normal guinea-pig serum (GPS) possesses a fl-like inhibitor which is closely related to elements of the classical complement pathway and that this inactivates influenza B virus but not influenza A virus. 0000-7423 1987 SGM

1136 F. YAMAMOTO AND OTHERS METHODS Virus, virus assay and cell culture. The following influenza virus strains were used in this study: B/Amagusa/64, B/Kagoshima/68, B/Osaka/70, B/Kanagawa/73, A/WSN (H1N1), A/FPV (H7N7) and A/PR8 (HIN1). The viruses were grown in the allantoic cavity of 1 l-day-old chick embryos at 34 C for 2 days or in MDCK (Madin- Darby canine kidney) cells for 2 to 3 days and stored at -80 C before use. Haemagglutinin (HA) titrations, NA assays and haemagglutination inhibition (HI) tests were performed as described by Maeno & Kilbourne (1970). Infectivity was titrated by plaque formation on MDCK cell cultures (Shibata et al., 1982 b) grown in Eagle's MEM containing 10~ foetal calf serum. Virus purification. Virus particles were pelleted by centrifugation at 20000 r.p.m, for 1 h in a SW27 rotor and resuspended in phosphate-buffered saline (PBS) ph 7-2. The virus suspension was layered on a 10 to 40~ linear sucrose gradient in PBS and centrifuged in a SW27 rotor for 1 h at 20000 r.p.m. The visible virus band was collected and diluted with PBS. The virus particles were sedimented by centrifugation at 20 000 r.p.m, for 1 h in a SW27 rotor. The pellet was resuspended in PBS. U.v. irradiation. One ml of purified virus suspension was placed in a 6 cm dish and exposed to u.v. radiation at a rate of 2.8 J/m 2.s with occasional shaking. Guinea-pig serum and its complement components. In the present study we used lyophilized whole GPS purchased from Kyokuto Seiyaku Ltd, Japan. Lyophilized GPS was dissolved in 1 ml of MEM or veronal-buffered saline (VS) ph 7-4. The stock virus preparations were clarified by centrifugation at a low speed, diluted 10-fold with MEM and mixed with an equal volume of GPS. At appropriate times during incubation the mixtures were diluted 100-fold with chilled MEM and examined for infectivity. Complement component CI (1200 CHs0 units) was isolated from fresh whole GPS (Gigli et al., 1976) and suspended in VS containing 1 mm-mgci2 and 0-15 mm- CaC12. GPS complement component C4 (1000 CHso units) was kindly supplied by Dr Moriyama, Osaka Medical College, Osaka, Japan. C4-deficient serum was collected from two guinea-pigs with hereditary C4 deficiency. Complement components and C4-deficient serum were stored at -80 C. In some experiments 10 m~l-mgcl 2, 10 mm-egta or 10 mm-edta were added to'gps. Treatment of GPS with zymosan. The treatment of GPS with zymosan was carried out according to the method of Inai et al. (1976). Briefly, zymosan (Sigma) was suspended in saline, boiled for 90 min, and then washed three times with saline. Activated zymosan thus obtained was suspended to 10 mg/ml in GPS containing Mg z+ and EGTA and incubated at 37 C for 15 min. Zymosan was removed by centrifugation at 3000 r.p.m, and CaCI 2 (0.025 M) was added to the GPS. Detection of GPS Clq antigen by the peroxidase-conjugated antibody technique. The procedure was carried out according to the method of Hawkes et al (1982). Briefly, a test sample (t ~tl) was dotted onto a nitrocellulose filter (Schleicher & Schuell) and dried at room temperature. The filter was soaked in 0.01 i-tris-buffered saline (TS) ph 7.4 containing 2~ bovine serum albumin (Sigma) overnight at 4 C, washed with TS, and then soaked in rabbit antiserum against GPS Clq (1:5000 dilution) for 30 min at room temperature. The antiserum was a gift from Dr Yonemasu, Nara Medical College, Nara, Japan. The filter was washed with TS containing 0-05 ~ Tween 20 (TS-T) and incubated with peroxidase-conjugated sheep anti-rabbit IgG (1:2000 dilution) for 30 min at room temperature. After washes with TS-T, the filter was soaked in the substrate solution containing 0.25~ 3,3'- diaminobenzidine-4hc1 and 0.01 ~ hydrogen peroxide in 0.1 i-tris-hc1 ph 7.4 at room temperature for 3 to 5 min and then washed with distilled water. A positive reaction was detected as a coloured dot against the white filter background. RESULTS Neutralization of influenza B virus by normal guinea-pig sera A 1 : 10 dilution of egg-grown influenza B/Kanagawa virus in MEM was mixed with an equal volume of normal GPS and incubated at 37 C for 45 min. GPS reduced the infectivity titre by more than 100-fold, but this inactivation did not occur when the mixtures were incubated at 0 C or when GPS had been heated at 56 C for 30 min (Table 1). Essentially the same results were obtained when MDCK-grown B/Kanagawa virus and other influenza B virus strains (B/Amagusa/63, B/Kagoshima/68 and B/Osaka/70) were used (data not shown). There was no significant difference in the inactivation effect among different batches of GPS (data not shown). These results suggest that neutralizing antibody is not responsible for the inactivating effect observed. Kinetic studies showed that inactivation of the virus by GPS reached a maximum level after incubation for 30 min (data not shown). If not stated otherwise, equal volumes of a 1 : 10 dilution of egg-grown influenza B/Kanagawa virus and GPS were incubated at 37 C for 30 min in all experiments below. A 1 : 10 dilution of GPS was a saturating dose for its

Inactivation of influenza virus 1137 Table 1. Inactivation of the infectivity of influenza B/Kanagawa by guinea-pig serum Reaction conditions r )" ~ Infectivity Reaction mixtures Temp. ( C) Time (rain) (p.f.u./ml) Virus + MEM 37 45 7.2 x 106 Virus + GPS 37 45 5-5 x 104 Virus + GPS (56 C, 30 min) 37 45 6.2 x 106 Virus + MEM 0 45 1.0 x 107 Virus + GPS 0 45 5.7 x 106 Virus + GPS (56 C, 30 min) 0 45 9.1 x 106 Table 2. Effect of trypsin treatment on inactivation activity of guinea-pig serum Trypsin concentration Infectivity (gg/ml) (p.f.u./ml)* 0 1.1 x 105 100 2.6 106 500 4.5 x 107 * Virus titre in the absence of GPS was 4.9 x 107. Table 3. Complement requirements for neutralization of influenza B/Kanagawa by guinea-pig serum Treatment of virus Infectivity (p.f.u./ml) Control 1.7 107 GPS 1.3 x l0 s GPS + EDTA 7.4 x 106 GPS + EDTA + Ca 2+ + Mg 2+ 2.0 x 10 ~ GPS + Mg 2+ + EGTA 2-0 x 107 GPS + Mg 2+ EGTA + Ca,-+ 2.9 x 10 s C4-deficient GPS 8.0 x 106 C4-deficient GPS + C4 < 1-0 x 10 ~ Zymosan-treated GPS 1.4 x 107 Zymosan-treated GPS + C4-deficient GPS 1.5 x 106 inactivation effect but 1 : 102 diluted GPS showed no appreciable effect. GPS was inactivated by trypsin (500 ~tg/ml; 36 C; 30 rain) followed by soybean trypsin inhibitor (500 gg/ml) (Table 2). Trypsin inhibitor alone did not inhibit virus infectivity or the inactivation activity of GPS (data not shown). Neutralization of influenza B virus by complement in GPS Influenza B virus was incubated with GPS containing EDTA (0.01 r~) and examined for residual infectivity (Table 3). Under these conditions GPS did not inactivate the virus, but addition of CaC12 (0.025 M) and MgCIz (0.02 ~) to EDTA-treated GPS restored its inactivation activity. When Mg 2 and EGTA were added to GPS instead of EDTA, the GPS also failed to inactivate the virus but its activity was restored by the addition of CaC12 (0.025 M) (Table 3). These results indicate that Ca 2 is required for inactivation by GPS. If complement were involved, the pathway of action would be the classical one, which is dependent on both Ca 2 and Mg 2 ions, rather than the alternative pathway, which requires only Mg 2 (Cooper, 1979, 1984). The initial activation of the classical complement pathway converts complement component C1 to its activated state in the presence of Ca z ; activated C1 then cleaves C4 and C2, the next two reacting components, to form C4b2a in a reaction depending on Mg 2. C4b2a subsequently cleaves C3 to its active form C3b which cleaves C5 to generate C5b. The membrane attack pathway is initiated by C5b (Cooper, 1979, 1984).

1138 F. YAMAMOTO AND OTHERS Fresh serum obtained from guinea-pigs with hereditary C4 deficiency showed no appreciable inactivation effect, but when C4 (final concentration 500 CHs0 units) was added to the C4- deficient GPS, the reconstituted serum significantly reduced the infectivity titre (Table 3). Zymosan, a polysaccharide isolated from yeast, is a potent activator of the alternative complement pathway. It has been found that treatment of serum with zymosan results in selective depletion of C3 and C5 (Inai et al., 1976; Fearon & Austen, 1977). This led us to investigate whether completion of the C2 step in the classical complement pathway was essential for inactivation of influenza B virus. When GPS was treated with zymosan (see Methods), the serum lost its inactivation activity (Table 3). But after the addition of an equal volume of C4-deficient GPS, zymosan-treated GPS exhibited inactivation activity, suggesting that the inactivation by GPS may require the completion of the C3 step in the classical pathway, or this together with the membrane attack pathway. The classical complement pathway is generally triggered by antibody complexed with specific antigens including viruses. Therefore the absorption of putative antibody with purified influenza B virus was carried out to eliminate the possible involvement of antibody. GPS containing EDTA (0.01 M) was incubated with u.v.-irradiated virus suspension (final concentration 3200 HAU) at room temperature for 30 min. The virus particles were sedimented by centrifugation in a SW50 rotor at 24000 r.p.m, for 1 h and then CaCI~ (0-025 M) and MgC12 (0.02 M) were added to the supernatant. Equal volumes of the supernatant and influenza B virus (2.0 x l0 T p.f.u./ml, 6 HAU) were incubated at 37 C for 30 rain and examined for infectivity. The treated GPS reduced the infectivity titre by more than 100-fold. A parallel experiment showed that when antiserum against influenza B virus was absorbed with the same amount of virus preparation, its HI titre was reduced from 6400 to less than 100. HA triggers the classical complement pathway A study was carried out to determine whether HA or NA on the viral envelope triggered the activation of the classical complement pathway. A previous paper (Shibata et al., 1982a) has shown that trypsin digestion of influenza B virus results in the selective removal from virions of the enzyme-active portion of the NA molecule. Trypsin (500 gg/ml) was added to purified MDCK-grown virus suspension and the mixture was incubated at 36 C for 30 rain. After the addition of soybean trypsin inhibitor (500 ~tg/ml), the virus suspension was centrifuged in a SW50 rotor at 24000 r.p.m, for 1 h. The virus pellet was resuspended in the original volume of PBS and tested for viral activities. More than 97 ~ of the virus-associated NA activity was lost but HA activity was comparable to that of untreated virus. Its infectivity was enhanced, possibly due to the cleavage of HA by trypsin. The treated virus sample was tested for its response to GPS, and was found to be inactivated to a level comparable to untreated virus, suggesting that inactivation by GPS is induced by the interaction of HA with complement. Effect of GPS and C1 on HA activity of influenza B/Kanagawa When influenza B virus was incubated with GPS at 0 C or 37 C for 30 rain, HA activity was completely inhibited (Table 4). The addition of EDTA (0.0t M) after incubation at 0 C restored HA activity, but neither HA activity nor infectivity was recovered by the addition of EDTA (0.01 M) after incubation at 37 C. Table 4 also shows that the first complement component C1 inhibited the HA activity, independently of the incubation temperature, but HA activity was readily recovered by the addition of EDTA (0.01 M) at either incubation temperature. C1 is composed of subunits C 1 q, C l r and C 1 s; the latter two subunits are bound to C1 q via a Ca 2 + ion, which recognizes the activator (Cooper, 1979, 1984). It is probable that the interaction of C1 with the virus sterically hindered binding of HA to the receptor on red cells, but it was not clear whether this was reversed because C 1 r and C 1 s dissociated from C 1 q on the addition of EDTA or because C1 was released. The former possibility was found to be more likely. Equal volumes (25 ~1) of purified virus suspension and C1 were incubated at 0 C for 30 min and layered on 20~ sucrose in VS containing 0-15 mm-cac12 and 1 mm-mgc12, followed by centrifugation at 24000 r.p.m, for 1 h in a SW50 rotor. The pellet was resuspended in the original volume of PBS and a sample (1 gl) of the suspension was dotted onto a nitrocellulose filter. The dot was incubated first

Inactivation of influenza virus 1139 Nil Fig. 1. Binding of GPS C 1 q to influenza B virus. Purified influenza B virus suspension (1280 HAU) was incubated with an equal volume (25 gl) of C1 (1200 CHs0) at 0 C for 30 rain (a). Immediately after the incubation 0.01 M-EDTA was added to another mixture (b). Each mixture was then layered on 20% sucrose [4-9 ml) in VS containing 0.15 mm-cac12 and 1 mn-mgcl2 and centrifuged at 24000 r.p.m, for 1 h in a SW50 rotor. Separate samples (50 gl) of virus suspension (c) and C1 (d) were also centrifuged under the same conditions. The pellets were resuspended in 50 gl PBS and samples (1 gl) of the suspensions were dotted onto nitrocellulose filters. The filters were soaked first in rabbit anti-gps C 1 q (1:5000 dilution), second in peroxidase-conjugated sheep anti-rabbit IgG (1:2000 dilution) and finally in substrate solution containing 0.25 % 3,3'-diaminobenzidine-4HC1 and 0.01% hydrogen peroxide in 0.1 M-Tris-HCI ph 7.4. The filters were washed with distilled water at 3 to 5 min after incubation with the substrate. A positive reaction was detected as a coloured dot against the white filter background. Table 4. Effect of guinea-pig serum or CI on the HA activity of influenza B/Kanagawa and A/WSN HA activity at Reaction mixture 0 C 37 C B/Kanagawa + MEM 64 128 B/Kanagawa + GPS <2 <2 B/Kanagawa + GPS + EDTA-~ 32 <2 A/WSN + MEM 128 ND* A/WSN + GPS 64 ND B/Kanagawa + MEM 64 64 B/Kanagawa + C1 4 4 B/Kanagawa + CI + EDTA t 32 32 A/WSN + MEM 128 ND A/WSN + C1 64 ND * ND, Not determined. t EDTA was added after the virus suspension had been incubated with GPS or C1 for 30 min. with rabbit antiserum against GPS Clq, second with peroxidase-conjugated sheep anti-rabbit IgG, and finally with the substrate. A coloured dot appeared on the white filter (Fig. 1 a). When EDTA was added to the mixture after the virus-c1 incubation, a coloured dot was also detected (Fig. 1 b). In parallel experiments the separate centrifugation of virus suspension and C 1 resulted in no development of colour (Fig. 1 c, d). These results also show that C 1 could bind to influenza B virus in the absence of antibody. Failure of GPS to inactivate influenza A viruses We also studied the effect of GPS on the infectivity of influenza A viruses. The viruses were diluted 1 : 10 with MEM, incubated with GPS at 37 C for 30 min, and examined for remaining infectivity. As shown in Table 5, GPS inactivated the influenza A viruses by less than threefold under conditions in which GPS reduced the infectivity of influenza B virus by more than 103- fold. Both GPS and C1 inhibited little of the HA activity of A/WSN when incubated at 0 C (Table 4). DISCUSSION The present study showed that influenza B/Kanagawa virus was efficiently neutralized by normal GPS lacking detectable antiviral antibody. This inactivation occurred only when virus and serum were incubated at 37 C, and the inhibitor was heat-labile and sensitive to trypsin digestion. These features suggest a fl inhibitor (Chu, 195l) and the inhibitor in GPS acting against influenza B virus was found to be the complement system. The complement system consists of 20 different protein components that must be activated to

1140 F. YAMAMOTO AND OTHERS Table 5. Comparison of sensitivity of influenza A and B viruses to guinea-pig serum Reaction mixture Infectivity (p.f.u./ml) A/WSN + MEM 3.4 106 A/WSN + GPS 1.5 x 106 A/FPV + MEM 1.0 x 106 A/FPV + GPS 9.0 x 105 A/PR8 + MEM 4"0 x [06 A/PR8 + GPS 1"7 x 106 B/Kanagawa + MEM 4.4 x 106 B/Kanagawa + GPS 2.5 x 103 produce their biological effects. There are two activation pathways, the classical complement pathway and the alternative pathway, and a terminal membrane attack pathway. The classical complement pathway depends on both Ca 2 and Mg 2 ions and is driven by a series of enzymes which are generally formed from inactive complement precursors C1, C4, C2 and C3 by limited proteolytic cleavage by other complement enzymes. The alternative complement pathway depends on Mg 2 and involves the interaction of C3, Factor B and Factor D in the absence of antibody, leading to formation of C3bBb (Cooper, 1979, 1984; Sissons, 1984). C4-deficient GPS had no ability to inactivate influenza B virus but it drastically inactivated the virus after the addition of C4, indicating that the inactivation activity depends on functions of C1 and C4. Together with the requirement for Ca 2 ions for this effect, these results indicate that the classical rather than the alternative complement pathway is responsible for inactivation by GPS. Zymosan is a potent activator of the alternative complement pathway and treatment of serum with it results in selective consumption of C3 and C5 (Inai et al., 1976; Fearon & Austen, 1977). Zymosan-treated GPS did not inactivate influenza B virus but the inactivation effect was exhibited by the addition of C4-deficient GPS, suggesting that completion of the C3 step in the classical complement pathway may be essential for inactivation. It remains to be determined whether the classical complement pathway is needed alone or together with the membrane attack pathway for inactivation by GPS. The classical complement pathway is generally triggered by antibody in complexes with viral antigen (Cooper, 1984). Recently Beebe et al. (1983) have demonstrated that human serum neutralizes influenza A/WSN via the classical complement pathway with the participation of trace antibody undetectable by HI tests. The presence of antibody seems less likely in our system, since (i) absorption of putative antibody with concentrated virus particles has no effect on the inactivation activity of GPS and (ii) isolated C1 alone binds to influenza B virus. Several lines of evidence have also demonstrated that the classical pathway is activated by the interaction of C1 with different substances including a polypeptide located on the external surface of virus, polyanions such as RNA or DNA, certain cellular membranes, and lipid A of lipopolysaccharides (Cooper, 1984). GPS C1 appears to interact directly with the HA and/or NA glycoproteins, leading to the activation of complement system, since they are located externally on the envelope of influenza B virus. When more than 97 % of NA activity was removed from the virions by trypsin digestion, the inactivation activity of G PS was not significantly affected, suggesting that interaction of HA with complement is sufficient to inactivate the virus. The carbohydrate moieties of HA and NA are determined by the host cells in which the virus is grown. Since MDCK cell- and egg-grown influenza B viruses were inactivated to a similar extent by GPS, it is likely that C1 interacts directly with the protein moiety of the HA glycoprotein to activate the classical complement pathway. C1 is composed of one molecule of Clq (mol. wt. 400000), two molecules of Clr (mol. wt. 190 000) and two molecules of C1 s (mol. wt. 88 000) and both C 1 r and C1 s associate with C 1 q via a Ca 2 ion, which recognizes the activator (Cooper, 1979). The binding of C1 to influenza B virus inhibited its HA activity, and the dissociation of Clr and Cls from virion-bound Clq by

Inactivation of influenza virus! 141 EDTA treatment restored HA activity; this suggests that binding of C1 to influenza B virus causes steric hindrance of the attachment of HA to the receptor on the red cell surface and that this event requires the cooperation of all three C1 subunits. Nevertheless, C1 was unable to prevent infection by influenza B virus, and neither did C4-deficient GPS. One explanation for this may be that saturation of the virus particle with C 1 leaves active HA exposed at one site on the virus surface, allowing binding of the virus to cell receptors but no haemagglutination. In contrast, GPS inhibited both HA activity and infectivity at 37 C and these activities were not restored by the addition of EDTA. Activation of the classical complement pathway may result in the deposition of complement proteins all over the virus surface, leading to steric hindrance of attachment of HA to the cell receptor, or virus destruction. Influenza A and B viruses resemble each other in many structural and biological properties, but the complement system of GPS did not inactivate influenza A virus. Further study of the response of GPS complement to influenza A and B viruses will provide more information as to the mechanism of activation of the classical complement pathway. The help of Mr Etsuo Iida with the supply of lyophilized guinea-pig serum is acknowledged. This work was supported by research grants from the Ministry of Education, Science and Culture of Japan, and in part from Takano Hospital. REFERENCES ANANTHANARYAN, R. & PANIKER, C. J. K. (160). Non-specific inhibitors of influenza viruses in normal sera. Bulletin of the World Health Organization 22, 409-420. BEEBE, D. P., SCHRIBER, R. D. & COOPER, N. R. (1983). Neutralization of influenza virus by normal human sera: mechanisms involving antibody and complement. Journal oflmmunology 130, 1317-1322. ehu, C. M. (1951). The action of normal mouse serum on influenza virus. Journal of General Microbiology 5, 739-757. COOPER, N. g. (1979). Humoral immunity to viruses. In Comprehensive Virology, vol. 15, pp. 123-164. Edited by H. Fraenkel-Conrat & R. R. Wagner. New York & London: Plenum Press. COOPER, iq. R. (1984). The role of the complement system in host defense against virus diseases. In Concepts in Viral Pathogenesis, pp. 20-25. Edited by A. L. Notkins & M. B. A. Oldstone. New York, Berlin, Heidelberg & Tokyo: Springer-Verlag. FEARON, D. Y. & AUSTEN, K. F. (1977). Activation of the alternative complement pathway due to resistance of zymosan-bound amplification convertase to endogenous regulatory mechanisms. Immunology 74, 1683-1687. FRANCIS, X., JR (1947). Dissociation of hemagglutinating and antibody measuring capacities of influenza virus. Journal of Experimental Medicine 85, 1-7. GIGLI, I., PORTER, R. R. & SIM, g. B. (1976). The unactivated form of the first component of human complement, C 1. Biochemical Journal 157, 541-548. HAWKES, R., NIDAY, E. & GORDON, J. (1982). A dot-immunobinding assay for monoclonal and other antibodies. Analytical Biochemistry 119, 142-147. INAI, S., NAGAKI, K., EBISU, S., KATO, K., KOTANI, S. & MISAKI, A. (1976). Activation of the alternative complement pathway by water-insoluble glucans of streptococcus mutants: the relation between their chemical structures and activating potencies. Journal of Immunology 117, 1256--1260. KRIZANOVA, O. & RATHOVA, V. (1969). Serum inhibitors of myxoviruses. Current Topics in Microbiology and Immunology 4"7, 125-151. MAENO, K. & KILBOURNE, E. D. (1970). Developmental sequence and intracellular site of synthesis of three structural protein antigens of influenza A2 virus. Journal of Virology 5, 153-164. SHIBATA, M., MAENO, K., TSURUMI, T., AOKI, H., NISHIYAMA, Y., ITO, Y., ISOMURA, S. & SUZUKI, S. (1982a). Role of viral glycoprotein in haemolysis by influenza B virus. Journal of General Virology 59, 183-186. SHIBATA, M., MAENO, K., ISOMURA, S., TSURUMI, T., AOKI, H. & SUZUKI, S. (1982b). Plaque formation by influenza B virus in a porcine kidney cell line. Microbiology and Immunology 26, 441 A4~. SHIMOJO, H., SUGIURA, A., AKAO, J. & ENOMOTO, C. (i958). Studies of a non-specific hemagglutination inhibitor of influenza A (Asian) 57 virus. Bulletin of the Institute of Public Health 7, 219-224. SISSONS, J. G. P. (1984). Antibody- and complement-dependent lysis of virus-infected ceils. In Concepts in Viral Pathogenesis, pp. 39-45. Edited by A. L. Notkins & M. B. A. Oldstone. New York, Berlin, Heidelberg & Tokyo: Springer-Verlag. SUGIURA, A., SHIMOJO, H. & ENOMOTO, C. (1961). Studies of the non-specific hemagglutination inhibitor against influenza A2 virus (gamma-inhibitor). II. Physicochemical properties of the gamma-inhibitor. Japanese Journal of Experimental Medicine 31, 159-168. (Received 27 August 1986)