Molecular Biology, Vol. 36, No. 3, 2002, pp. 429–435. Translated from Molekulyarnaya Biologiya, Vol. 36, No. 3, 2002, pp. 542–549. Original Russian Text Copyright © 2002 by Gambaryan, Yamnikova, Lvov, Robertson, Webster, Matrosovich.
MISCELLANEOUS UDC 578.832
Differences in Receptor Specificity between the Influenza Viruses of Duck, Chicken, and Human A. S. Gambaryan1, S. S. Yamnikova2, D. K. Lvov2, J. S. Robertson3, R. G. Webster4, and M. N. Matrosovich1 1
Chumakov Institute of Poliomyelitis and Viral Encephalitides, Russian Academy of Medical Sciences, Moscow Region, 142782 Russia; E-mail:
[email protected] 2 Ivanovsky Institute of Virology, Moscow, 123098 Russia 3 National Institute for Biological Standards and Control, Herts EN6 3QG, UK; 4 Department of Virology and Molecular Biology, St. Jude Children’s Research Hospital, Memphis, Tennessee, 38105-2794, USA Received June 14, 2001
Abstract—The affinity of the duck, chicken, and human influenza viruses to the host cell sialosides was determined, and considerable distinctions between duck and chicken viruses were found. Duck viruses bind to a wide range of sialosides, including the short-stem gangliosides. Most of the chicken viruses, like human ones, lose the ability to bind these gangliosides, which strictly correlates with the appearance of carbohydrate at position 158–160. The affinity of the chicken viruses to sialoglycoconjugates of chicken intestine as well as chicken, monkey, and human respiratory epithelial cells exceeds that of the duck viruses. The human influenza viruses have high affinity to the same cells but do not bind at all to the duck epithelial cell. This testifies to the absence of 6'-sialylgalactose residues from the duck cells, in contrast to chicken and monkey cells. The alteration of the receptor specificity of chicken viruses in comparison with duck ones results in the similarity of the patterns of accessible cells for chicken and human influenza viruses. This may be the cause of the appearance of the line of H9N2 viruses from Hong Kong live bird markets with receptor specificity similar to that of H3N2 human viruses, and of the ability of H5N1 and H9N2 chicken influenza viruses to infect humans. Key words: influenza viruses, evolution, species barrier, receptors, gangliosides
INTRODUCTION Human influenza viruses replicate in the cells of the respiratory epithelium [1], whereas influenza viruses of wild aquatic birds predominantly replicate in the cells lining the intestinal tract [2]. Avian influenza viruses, as a rule, do not replicate efficiently in humans. For example, high doses of virus were found to be required for the replication of avian influenza strains in volunteers even at a limited level [3], and no cases of influenza virus infections were documented in workers exposed to highly pathogenic avian viruses during the 1985 outbreak in poultry in the USA [4].
influenza viruses. The human strain was shown to bind only to species with 10 or more sugars, while the duck strain bound to a wide range of gangliosides including 3–5-unit ones [10].
A factor restricting the host range of influenza viruses is their limited receptor specificity. It has been shown that influenza A viruses isolated from avian species preferentially bind to 3'-sialylgalactose (3'SG) terminated sugar chains (Neu5Acα(2-3)Galβ), while closely related human viruses reveal a higher binding affinity towards the 6'SG-terminated structures (Neu5Acα(2-6)Galβ) [5–9].
Until recently it has been supposed that 6'SG specificity is obligate for influenza virus propagation in the human respiratory tract cells [11]. Nevertheless, chicken H5N1 viruses were isolated from a human in 1997 [12–16] though they retained the original 3'SG specificity [17]. Thus one can suppose that the afflicted human cells have 3'SG receptors with high enough affinity for the avian virus. Does this mean that any avian influenza viruses can propagate in the human respiratory tract, or the receptor properties of the chicken viruses change so that the human cells become more susceptible for them than for the classical duck viruses? There are data that transmission of the virus from duck to chicken is attended by its rapid evolution comparable to that upon transmission to mammals [18].
Marked distinctions were demonstrated in the recognition of gangliosides between duck and human
Such alterations include deletion in the stalk of the neuraminidase and additional glycosylation sites in
0026-8933/02/3603-0429$27.00 © 2002 MAIK “Nauka /Interperiodica”
430
GAMBARYAN et al.
the hemagglutinin receptor pocket [17]. It can thus be suggested that the virus thereby adapts to receptors with a longer carbohydrate stem. To check this suggestion, we determined the affinity of a number of avian and human influenza viruses to sialoglycoconjugates of the plasma membranes of a number of host cells and to gangliosides therefrom differing in the sugar stem length. EXPERIMENTAL Viruses. H5N1 was from the collection of St. Jude Hospital, human species from the National Institute for Biological Standards and Control, others from the collection of the Ivanovsky Institute of Virology. The viruses were grown either in 10-day-old embryonated chicken eggs or in MDCK tissue culture as described before [9] and stored in 60% glycerol in 0.1 M NaCl, 0.02 M Tris-HCl with 0.02% sodium azide at –20°ë. Tissues. Freshly killed adult chickens (Gallus gallus) and mallard ducks (Anas platyrhynchos) were purchased from a live bird market in Moscow; intestine, trachea, and lung were taken immediately. African green monkey (Cercopithecus aethiops) trachea and lung were obtained in the Chumakov Institute of Poliomyelitis and Viral Encephalitides. Plasma membranes. Epithelial cells were scratched off with a spatula, washed in PBS and freed from red blood cells by low-speed centrifugation in 50% Percoll. Plasma membranes were obtained by hypotonic shock and homogenization, debris removal at 1000 g, and pelleting at 40,000 g for 2 h. Total gangliosides. Membranes were suspended in methanol and a double volume of chloroform was added. The suspension was incubated 1 h at 45°C, centrifuged, and the pellet was re-extracted with 80% ethanol for 2 h at 45°C [19]. The extracts were evaporated, dissolved in chloroform–methanol (2:1) and pooled. Then one fifth volume of water was added, the mix was intensely shaken, kept for 1 h at 45°C, and centrifuged. The top phase was collected, dried, and dissolved in 50% ethanol [20]. Thin-layer chromatography. Analytical TLC of total gangliosides was performed by using HPTLC silica gel 60 plates with a concentrating zone (Merck, Germany) as a mobile phase (Portner et al., 1993). (Svennerholm, 1957). Chromatographs were scanned and processed using Adobe Photoshop 5.0 software. Total gangliosides (ca. 5 nmol in 2–3 µl) were fractionated by preparative TLC on Polygram Sil G plates (Macherey-Nagel, Germany) in chloroform/methanol/20 mM CaCl2 (120:85:20). Resorcinol was used to stain sialic acids [21]. In parallel, the bands were visualized by iodine staining, and all visible spots and sections between spots were scratched off and extracted with 50% methanol at 37°C.
Solid phase binding assay. Aliquots (0.1 ml) of the sonicated glycoprotein or membrane preparation (10 µg protein per ml PBS) were incubated in the wells of the EIA polystyrene 96-well microplate for 2 h; 20 µl of the solutions of gangliosides in 80% methanol containing egg lecithin (8 µg/ml) and cholesterol (4 µg/ml) was added per well and evaporated. The optimal ganglioside concentration was determined empirically. Positive control wells contained fetuin, negative control wells contained only lecithin and cholesterol. Plates were washed with water and blocked with BSA for an hour. Then 0.05 ml of serial two-fold dilutions of the virus in PBS with 2 mg/ml BSA and 2 µM amino-GN (neuraminidase inhibitor) were incubated in the wells for 2 h at 4°C. After washing, 0.05 ml/well of fetuin-HRP conjugate in 0.02 M Tris-HCl pH 7.4 0.15 M NaCl, 0.01% Tween-20, 1 mg/ml BSA, 2 µM amino-GN was added for 30 min at 4°C. After washing, o-phenylenediamine was added and the absorbance at 492 nm was measured; the data were converted to Scatchard plots. As a positive control of the maximal virus binding (K+ control), noncoated microplates were used, and the assay followed the same protocol, except that the blocking step was omitted and no BSA or Tween was present in the incubation mixture. As a result, the virus was allowed to adsorb to the plastic nonspecifically, providing the measure of the maximal possible binding as dependent on the virus concentration. To compare different viruses, their concentrations were chosen so that the slopes coincided. The slope of the K+ Scatchard plot was taken as 100%, and the relative affinity to plasma membranes and to sialosides was estimated with respect to this control. Treatment of coated plates with neuraminidase and trypsin. About 10–5 un. Vibrio cholerae neuraminidase or 0.02 mg trypsin (both from Serva) in 0.2 ml of 0.05 M acetate buffer pH 5.5 with 2 mM MgCl2, 10 mM CaCl2, 1 mg/ml BSA were added per well and incubated overnight at 37°C in a humid chamber. The plates were washed and used for further experiments. RESULTS AND DISCUSSION To assess the alteration of receptor specificity of influenza viruses after transmission to a new host, we investigated the binding of viruses to some host cell sialosides. The virus attachment to cell plasma membranes, sialylglycoproteins and gangliosides on a solid phase was assayed using the microplate adsorption method [22]. For comparison, we used the natural sialoglycoprotein fetuin containing both 3'SG and 6'SLN groups [23], and bovine brain gangliosides (BBG): a GM1, GD1a, GD1b, GT1b mixture [24]. Neuraminidase-treated samples were used as background control. Such treatment entirely prevented the MOLECULAR BIOLOGY
Vol. 36
No. 3
2002
RECEPTOR SPECIFICITY OF INFLUENZA VIRUSES
(b)
(c)
(d)
K+ Fet G Fet-Tr BBG-Tr
Ä 492/ë
(a)
431
ä DkI ChI DkI-Tr ChI-Tr
Ä 492 Fig. 1. Scatchard plots for the binding of duck (a, b) and chicken (c, d) viruses with model sialosides (a, c) and host cell receptors (b, d). Designations as in Table 1; positive control (K+) shown with solid line.
binding of viruses to all used specimens. For determination of the type of sialoside (glycoprotein or glycolipid) that bind with the viruses, the samples were treated with trypsin. The titration data were presented as Scatchard plots. The effect of trypsin treatment on the chicken and duck virus binding to fetuin and BBG is shown in the upper part of Fig. 1. The binding to fetuin was entirely prevented, and binding to BBG was not affected. The number of binding sites for fetuin-coated wells is at the upper limit, because the point of intersection of the Scatchard plot with the abscissa axis coincides with that in the positive control; but the binding affinity is lower (the slope is about three times smaller). It is also seen that the duck virus binds to BBG as well as to fetuin, but the chicken virus binds much more poorly. In the lower part of Fig. 1, analogous Scatchard plots for binding of these viruses to plasma membranes of duck intestine and chicken trachea epitelial cells are shown. The duck virus plots are parallel and intersect the abscissa axis close to the positive control MOLECULAR BIOLOGY
Vol. 36
No. 3
2002
plot. This suggests that the number of virus binding sites is large, and their affinity is much the same. The number of binding sites is only slightly reduced by the trypsin treatment, which indicates that gangliosides are the prevalent receptors for the duck virus. The chicken virus plots are different. The points of intersection with the abscissa axis are at the left side of the positive control point, which indicates that the number of available receptors for the chicken virus is less than for the duck virus. The trypsin treatment of duck intestine preparation slightly reduces the binding of virus, but the treatment of chicken trachea preparation reduces the number of binding sites and increases the slope, suggesting that a limited number of nonprotein receptors bind the virus with high affinity. Thus one can conclude that the receptor determinant for the two viruses are dissimilar. The binding of a number of duck, chicken, and human viruses to the epithelial cells of duck and chicken intestine, chicken and monkey trachea, monkey nasal cavity, and human throat were estimated
432
GAMBARYAN et al.
Table 1. Relative affinity of viruses from different hosts to total cell sialoglycoconjugates and gangliosides Sialoglycoconjugates Viruses
BBG
Duck NewJersey/1580/78 HongKong/308/78 Chicken HongKong/786/97 FPV/Rostok/34 Human A/Chr/157/83M B/NIB/48/90M
Fet
DkI
ChI
ChT
MkT
HuT
–
Tr
–
Tr
–
Tr
–
Tr
–
Tr
–
Tr
–
Tr
(H2N3) (H5N1)
30 30
30 30
20 30
0 0
30 30
30 30
30 30
30 30
30 40
30 40
30 40
30 40
30 40
30 40
(H5N1) (H7N1)
