|Home | About | Journals | Submit | Contact Us | Français|
Quail are thought to serve as intermediate hosts of influenza A viruses between aquatic birds and terrestrial birds, such as chickens, due to their high susceptibility to aquatic-bird viruses, which then adapt to replicate efficiently in their new hosts. However, does replication of aquatic-bird influenza viruses in quail similarly result in their efficient replication in humans? Using sialic acid-galactose linkage-specific lectins, we found both avian (sialic acid-α2-3-galactose [Siaα2-3Gal] linkages on sialyloligosaccharides)- and human (Siaα2-6Gal)-type receptors on the tracheal cells of quail, consistent with previous reports. We also passaged a duck H3N2 virus in quail 19 times. Sequence analysis revealed that eight mutations accumulated in hemagglutinin (HA) during these passages. Interestingly, many of the altered HA amino acids found in the adapted virus are present in human seasonal viruses, but not in duck viruses. We also found that stepwise stalk deletion of neuraminidase occurred during passages, resulting in reduced neuraminidase function. Despite some hemagglutinin mutations near the receptor binding pocket, appreciable changes in receptor specificity were not detected. However, reverse-genetics-generated viruses that possessed the hemagglutinin and neuraminidase of the quail-passaged virus replicated significantly better than the virus possessing the parent HA and neuraminidase in normal human bronchial epithelial cells, whereas no significant difference in replication between the two viruses was observed in duck cells. Further, the quail-passaged but not the original duck virus replicated in human bronchial epithelial cells. These data indicate that quail can serve as intermediate hosts for aquatic-bird influenza viruses to be transmitted to humans.
The pandemic influenza viruses of 1957 and 1968 arose by reassortment between human and avian viruses. A prerequisite of reassortant generation is coinfection of a single host with at least two viruses, followed by replication of the reassortant viruses in the host. However, human viruses do not replicate well in birds, and similarly, duck viruses grow poorly in humans (13). This host range restriction is partly attributable to receptor specificity; most avian influenza A viruses preferentially bind to the sialic acid-α2-3-galactose (Siaα2-3Gal) linkage on sialyloligosaccharides, whereas human viruses preferentially bind to Siaα2-6Gal (25). By the same token, epithelial cells in the upper respiratory tract of humans mainly contain receptors for human influenza viruses (i.e., Siaα2-6Gal) (36), whereas the cells in the duck intestine (the major replication site for duck viruses) mainly contain avian virus receptors (i.e., Siaα2-3Gal) (27). Pigs are thought to be an intermediate host for the generation of reassortants between human and avian viruses, since they support the replication of both avian and human influenza viruses (19) and both receptor types are present on the epithelial cells of the pig trachea (1, 15, 30). Indeed, the emergence of pandemic (H1N1) 2009 influenza demonstrated that pandemic viruses could be generated in pigs. However, some avian influenza viruses, such as H5N1, H7N7, and H9N2 viruses, are occasionally transmitted from terrestrial poultry to humans, suggesting another mechanism for the generation of pandemic influenza virus strains. The identification of other potential intermediate hosts is of great importance to mitigate the public health impact of the emergence of influenza viruses with pandemic potential.
Previous studies have reported that the receptor binding properties of viruses circulating in terrestrial poultry differ from those of viruses circulating in aquatic birds (9, 24, 26, 43); H5N1 viruses isolated from chickens have a substantially lower affinity for Siaα2-3Gal than do aquatic-bird viruses (9, 26), and this reduced affinity is similar to that of human influenza viruses, although H5N1 chicken isolates do not exhibit Siaα2-6Gal specificity. Moreover, H5N1 chicken isolates with reduced receptor binding affinity have an additional glycosylation site in the globular head region of their hemagglutinin (HA) and a deletion in their neuraminidase (NA) stalk region—molecular features that are found in the glycoproteins of human viruses (26). Strikingly, H9N2 and H7N2 viruses isolated from terrestrial poultry, but not from aquatic birds, show receptor specificity similar to that of human isolates (2, 26, 35, 43).
Among terrestrial poultry, quail are considered an intermediate host for duck viruses. Makarova et al. and Perez et al. (23, 33) showed that quail are more susceptible than chickens to 14 of the 15 HA subtypes of aquatic-bird viruses and that the amino acid pattern of quail H9 viruses is intermediate between those of duck and chicken viruses at the seven positions on HA reported to be responsible for the adaptation of H9 viruses in terrestrial poultry (33). Further, this study showed that serial passage of duck H9N2 viruses in quail enables the viruses to replicate and be transmitted efficiently in chickens (14, 38, 39). They also found that both Siaα2-3Gal and Siaα2-6Gal glycoconjugates were present in the quail trachea and intestine and that avian and human viruses bound to the epithelial cells at these sites (10, 20, 42). These observations indicate that the changes that occur in the HA of duck viruses during their adaptation to quail may allow duck influenza viruses to cross the species barrier and infect terrestrial poultry, such as chickens. However, whether this adaptation to quail allows duck viruses to replicate efficiently in human cells remains unknown. To better understand the role of quail in this context, we passaged a duck H3N2 influenza virus in quail and compared the receptor specificities, as well as the replicative abilities in cell culture and human respiratory tissues, of the quail-passaged viruses.
Madin-Darby Canine Kidney (MDCK) cells and chicken embryo fibroblasts (CEF) were maintained in Eagle's minimal essential medium (EMEM) containing 5% (vol/vol) newborn calf serum (NCS). Duck embryo fibroblasts (DEF) were maintained in EMEM containing unheated 5% (vol/vol) NCS according to the manufacturer's instructions. Normal human bronchial epithelial (NHBE) cells were obtained from Lonza (Walkersville, MD) and maintained in serum-free and hormone-supplemented bronchial epithelial growth medium (SABM; Cambrex) containing bovine pituitary extract (BPE) (30 μg/ml), hydrocortisone (0.5 μg/ml), human epidermal growth factor (hEGF) (0.5 ng/ml), epinephrine (0.5 μg/ml), transferrin (10 μg/ml), insulin (5 μg/ml), triiodothyronine (6.5 ng/ml), bovine serum albumin–fatty-acid free (BSA-FAF; 50 μg/ml), retinoic acid (RA) (0.1 ng/ml), gentamicin (30 μg/ml), and amphotericin B (15 ng/ml). All cells were incubated at 37°C with 5% CO2. All viruses used in this study were propagated in MDCK cells in MEM supplemented with 0.3% (vol/vol) BSA and were stored at −80°C until use. The viruses used in this study were influenza A/duck/Mongolia/301/01 (H3N2) virus, A/Kawasaki/173/01 (H1N1) virus, and A/duck/Vietnam/5001/04 (H5N1) virus. All experiments with live H5N1 viruses were performed in biosafety level 3 (BSL3) containment laboratories at the University of Tokyo (Tokyo, Japan), which are approved for such use by the Ministry of Agriculture, Forestry and Fisheries, Japan.
Five-week-old Japanese specific-pathogen-free quail (Nisseiken Co., Ltd., Tokyo, Japan; n = 3) were inoculated intranasally (i.n.) and intratracheally (i.t.) with 0.5 ml of allantoic fluid containing 108.7 50% egg infectious doses (EID50) of virus. They were euthanized 3 days postinfection, and their organs (nasal turbinate, trachea, lung, and colon) were collected. The virus was serially passaged in quail (three quail per passage) 19 times with 0.5 ml of 10% (wt/vol) pooled organ homogenate (i.e., nasal turbinate, trachea, lung, and colon) every 3 days. The titer of the virus in each organ was recorded as the median 50% tissue culture infective dose (TCID50), that is, the inverse of the dilution that resulted in the cytopathic effect (CPE) in 50% of wells, using MDCK cells.
Nasal turbinates, trachea, lungs, and colon were collected, embedded in Tissue-Tek O.C.T. compound (Sakura Finetechnical Co., Ltd., Tokyo, Japan), and frozen on a dry ice block. Sections of each tissue (8 μm) were cut with a cryostat (CM 3050S; Leica Microsystems, Nussloch, Germany), air dried, and fixed for 10 min with cold acetone before being lectin stained. To detect sialyloligosaccharides reactive with Siaα2-3Gal- or Siaα2-6Gal-specific lectins, we incubated the sections with 250 μl of either fluorescein-labeled Sambucus nigra lectin (1:100; FL-1301; Vector Laboratories, Burlingame, CA) or biotinylated Maackia amurensis lectin II (1:100 dilution; B-1265; Vector Laboratories) overnight at room temperature. After being washed three times with Tris-buffered saline (TBS) (pH 7.6), the sections were incubated with Alexa Fluor 594-conjugated streptavidin (1:100 dilution; P-11227; Molecular Probes, Inc., Eugene, OR) for 1.5 h at room temperature. The sections were then counterstained with 4′,6-diamino-2-phenylindole dihydrochloride (Cellestain DAPI solution; 340-07971; Dojindo Molecular Technologies, Inc., Kumamoto, Japan). They were then washed three more times with TBS, covered with a coverslip, and observed under a fluorescence microscope (Eclipse TE300 with the fluorescence equipment mercury set; Nikon Co., Tokyo, Japan). Photographs were taken using a digital microscope camera (Olympus DP70; Olympus Optical Co., Ltd., Tokyo, Japan).
Viral RNA was extracted from virus in cell culture (MDCK) fluids by use of a commercial kit (Isogen LS; Nippon Gene, Tokyo, Japan) according to the manufacturer's instructions and converted to cDNAs by using primers containing the consensus sequences of the 3-prime ends of the RNA segments for HA,d NA, and reverse transcriptase (Superscript III; Invitrogen). The resultant cDNA products were used to amplify the HA and NA genes by a standard PCR method (Proof Start DNA polymerase; Qiagen). The purified PCR products were ligated into the pCR-Blunt II-Topo vector (Invitrogen) and used to transform Top10 cells (Invitrogen). Positive clones were cultured in Luria broth containing 50 mg/liter kanamycin and incubated overnight at 37°C in a shaking incubator. The bacterial culture was then pelleted by centrifugation, and the plasmid DNA was extracted for sequencing with a MagExtractor plasmid system (Toyobo, Osaka, Japan). The HA and NA gene sequences were analyzed with an Applied Biosystems (Foster City, CA) 3100 Auto Sequencer by use of cycle-sequencing dye terminator chemistry (Perkin Elmer, Boston, MA). The HA and NA genes of at least five cDNA clones were sequenced, and the other major sequences were determined for each sample. Primer sequences are available upon request.
Glycopolymers composed of poly-α-l-glutamic acid backbones with 5-N-acetylneuraminic acid linked to galactose through either an α-2-3 or an α-2-6 bond (Neu5Acα2-3LacNAcb-pAP and Neu5Acα2-6LacNAcb-pAP) were chemoenzymatically synthesized as described elsewhere (40). Virus suspension (200 HA units/ml) diluted in ice-cold PBS was used to coat 96-well polystyrene microplates (F96 Cert.Maxi Sorp Nunc-Immuno Plate; Nunc, Denmark), which were then incubated for 5 h at 4°C (on ice). As a control, wells without virus were also incubated. Unbound virus was removed by washing the wells three times with ice-cold PBS. The wells were then blocked by incubating them at 4°C overnight with 300 μl of PBS containing 0.001% Tween 20 (TPBS). The virus-coated wells were then washed a further three times with ice-cold PBS before 25 μl of horseradish peroxidase (HRP)-conjugated bovine fetuin (which possesses both Neu5Acα2-3Gal and Neu5Acα2-6Gal) diluted in TPBS (1:2,000) was added. Then, 25 μl of serially diluted sialylglycoconjugated polymers was added, and the plates were incubated at 4°C for 2 h. After being washed five times with ice-cold PBS, the plates were incubated with 100 μl of substrate solution (0.4 mg/ml of O-phenylenediamine, 0.01% H2O2 in 50 mM citrate-phosphate buffer, pH 5.5) at room temperature for 10 to 20 min. To stop the reaction, 50 μl of 0.1 N H2SO4 was added to each well. The extent of inhibition of fetuin binding to virions with sialylglycoconjugate polymers was determined by measuring the absorbance at 490 nm.
Viruses were grown in MDCK cells and ultracentrifuged at 25,000 rpm for 2 h at 4°C and then laid over a cushion of 25% sucrose in PBS. Virus stocks were aliquoted and stored at −80°C. Virus concentrations were determined by use of an HA assay with 0.5% (vol/vol) turkey red blood cells (RBCs). Custom microarray slides were printed for the Centers for Disease Control and Prevention (CDC) using the CFG glycan library (CDC version 1 slides; see Table S1 in the supplemental material for the glycans) as described previously (4). Virus preparations were thawed and suspended in PBS supplemented with 3% (wt/vol) BSA to an HA titer of 128, established to be optimal for glycan array analyses. Virus suspensions were supplemented with 10 nM zanamivir, overlaid on the printed region of the slides, and then incubated with gentle agitation in a closed container for 1 h at 4°C. Unbound virus was then eluted with brief rinses in PBS. The slides were immediately incubated with a mouse anti-Aichi/68HA monoclonal antibody (30 min), a biotinylated anti-mouse-IgG antibody (30 min), and a streptavidin-Alexa Fluor 635 conjugate (30 min) (Invitrogen, Carlsbad, CA) with brief PBS washes between incubations. After the final PBS wash, the slides were briefly rinsed in deionized water, dried under a gentle stream of air, and heat treated at 75°C for 1 h prior to imaging. Fluorescence intensities were captured by using a ProScanArray HT (PerkinElmer, Waltham, MA). ImaGene 8 software (BioDiscovery, El Segundo, CA) was used for image analyses. Data were processed in MS Excel to group similar sialoglycans and to generate a simplified chart. Three independent experiments were performed for each virus.
Reassortant viruses were generated by using the plasmid-based reverse-genetics system described previously (31). Viral RNA was extracted from virus in cell culture (MDCK) fluids by use of a commercial kit (Isogen LS; Nippon Gene, Tokyo, Japan) according to the manufacturer's instructions and converted to cDNAs by using primers containing the consensus sequences of the 3-prime ends of the RNA segments for HA, NA, and reverse transcriptase (Superscript III; Invitrogen). The resultant cDNA products were then used to amplify the HA and NA genes by a standard PCR method (Proof Start DNA polymerase; Qiagen). The cDNAs were cloned into a plasmid under the control of the human polymerase I promoter and the mouse RNA polymerase I terminator (referred to as PolI plasmids). Viruses possessing the HA and NA genes of plaque-purified parent virus or quail-passaged viruses (P19T clone 2 or clone 4), which have the internal genes of the A/whistling swan/Shimane/499/83 (H5N3) virus, were generated by using reverse genetics and are referred to as WT(HA,NA)-RG, P19Tcl2(HA,NA)-RG, and P19Tcl4(HA,NA)-RG. Viruses possessing the internal genes of the A/Yokohama/2017/03 (H3N2) virus were also generated for the growth curve in NHBE cells and are referred to as WT(HA,NA)-human-RG, P19Tcl2(HA,NA)-human-RG, and P19Tcl4(HA,NA)-human-RG. WT(HA)-P19Tcl2(NA)-RG and WT(HA)-P19Tcl4(NA)-RG were also generated for the virus elution assay.
The ability of NA to elute virus bound to erythrocytes was assessed as follows. Fifty microliters of 2-fold dilutions of virus containing HA titers of 1:128 was incubated with 50 μl of 0.5% (vol/vol) chicken erythrocytes or 1% (vol/vol) guinea pig erythrocytes in microtiter plates at 4°C for 1 h. The microtiter plates were then stored at 37°C, and the reduction in HA titers was recorded periodically for 12 h. Calcium saline (6.8 mM CaCl2-154 mM NaCl in 20 mM borate buffer, pH 7.2) was used as a diluent.
DEF, CEF, MDCK cells, and NHBE cells were infected with virus at a multiplicity of infection (MOI) of 0.001. After adsorption for 1 h, medium containing virus was removed, and the cells were overlaid with EMEM containing 0.3% BSA and 0.5 μg/ml of trypsin for DEF, CEF, and MDCK cells or serum-free, hormone-supplemented SABM for NHBE cells. The plates were then incubated at 37°C in 5% CO2. At various times postinfection, the virus titers in the cell culture supernatant were determined by using plaque assays with MDCK cells.
Fresh, surgically removed normal human lung specimens containing bronchi and alveoli were cut into ~0.5-cm3 cubes, washed with culture medium (F-12K nutrient mixture with 15% fetal calf serum [FCS], l-glutamine, and antibiotics), and incubated with virus (~108 PFU/ml) at 37°C in the medium. Twelve hours postinfection, the tissue blocks were fixed in 10% neutral-buffered formalin and processed for routine paraffin embedding and immunohistochemical analysis with a mouse anti-influenza A virus nucleoprotein antibody. The reactions were visualized by using a two-step dextran polymer system (Dako) and 3,3′-diamino benzidine (DAB). These experiments were performed with tissues from three patients. Because the results for all three were comparable, the findings from only one set of specimens are shown.
The glycan microarray data presented here are available online through the Consortium for Functional Glycomics website (http://www.functionalglycomics.org). Resource request identifiers are cfg_rRequest_1336/95 (parent virus), cfg_rRequest_1336/96 (P19T), cfg_rRequest_1336/97 (clone 2), and cfg_rRequest_1336/98 (clone 4).
The distribution of Siaα2-6Gal and Siaα2-3Gal on the epithelial cells of the nasal turbinates, trachea, lungs (parabronchus and mesobronchus), and colon was examined by using sialic acid-galactose linkage-specific lectins. In agreement with previous reports (20, 42), S. nigra lectin, specific for Siaα2-6Gal glycoconjugates, strongly reacted with the epithelial cells in the trachea (Fig. 1C). In contrast, M. amurensis lectin II, specific for Siaα2-3Gal, bound poorly to epithelial cells of the respiratory organs (Fig. 1); this is in contrast to previous reports (20, 42) in which M. amurensis lectin reacted strongly with the epithelial cells in quail trachea. This discrepancy may originate from a difference in the source of the Siaα2-3Gal linkage-specific lectins used. In addition to previous findings (10, 20, 42), we found that the S. nigra lectin-reacting Siaα2-6Gal glycans were abundant in the nasal turbinates and mesobronchus (Fig. 1). Interestingly, the parabronchi (Fig. 1E) lacked glycans that reacted with the S. nigra lectin. Of note, the secretory cells of these respiratory organs contained either Siaα2-3Gal or Siaα2-6Gal or both, as shown by their reactivity with the S. nigra lectin and/or M. amurensis lectin II (Fig. 1B and C, yellow-orange and arrows and arrowheads). The biological significance of this finding is unknown. In the colon, abundant secretory substances in the mucosal epithelium reacted strongly with S. nigra lectin, in accordance with a previous report (10), whereas in the deep crypts, the secretory epithelia contained both Siaα2-3Gal and Siaα2-6Gal, as indicated by the dual binding to both lectins (Fig. 1F, yellow). This pattern differed from duck intestine, in which M. amurensis lectin II strongly bound to the mucosal surfaces of the crypts (Fig. 1H, red).
To test the possibility that quail serve as an intermediate host for aquatic-bird influenza A viruses to improve their replication in human cells, A/duck/Mongolia/301/01 (H3N2) was serially passaged in quail. During these passages, viral loads in various organs (nasal cavity, trachea, lung, and colon) were determined by TCID50 titration on MDCK cells. Although the duck virus did not replicate well initially, it adapted to replicate more efficiently after repeated passages in quail, particularly in the respiratory organs (Fig. 2). Virus titers in the nasal turbinate, trachea, and lung after 19 passages were at least 2, 3, and 4 log units, respectively, higher than those after the initial passage, indicating viral adaptation in quail. Interestingly, virus titers in the colon remained relatively low even after 19 passages in quail.
Since HA is known to play a major role in viral adaptation to a new host (37) and the balance between HA receptor binding and NA receptor-destroying sialidase activity is important for efficient viral replication (29), we sequenced the HA and NA genes of five plaque-purified viruses from the trachea of quail during the repeated passages (Table 1). For HA, amino acid changes were found in viruses at different passage numbers, and some of these changes were maintained in viruses on further passage. After 19 passages, eight amino acid changes had accumulated in HA. To speculate about the biological significance of these amino acid changes, we mapped them on the three-dimensional structure of the H3 HA molecule (Fig. 3A and B).
As shown in Fig. 3B, some of the mutations occurred on the periphery of the receptor binding pockets (N137S and S227P). The lower left and right edges of the receptor binding site are formed by two surface loops: the 220 loop (residues 225 to 228) and the 130 loop (residues 135 to 138) form the lower left and right edges of the receptor binding site, respectively, while the α-helix (190 helix), which comprises residues 187 to 194, forms the top edge of the site (12). Given that residues 137 and 227 are located at or near the edge of the receptor binding pocket, the mutations found in these residues may have been positively selected during adaptation in quail. Residues 159 and 188 are located on the globular head of the HA1 portion of HA; therefore, the N159D and N188D changes may also affect HA receptor binding. Residues 325 and 333 are located at positions −5 and +4 from the HA1-HA2 cleavage site; their biological significance is unknown. Of note, these amino acid substitutions were not predicted to alter HA glycosylation.
Sequence analysis of the NA genes revealed a 17-amino-acid deletion (residues 67 to 83) in the NA stalk at passage 7 (Table 1 and Fig. 3C). An additional 7-amino-acid deletion (residues 39 to 45) in the stalk was detected in this virus at passage 15. Mutations at positions 46 and 215 were also detected in viruses isolated at passages 15 and 19; residue 46 is located on the stalk, and residue 215 is on the outermost β-sheet of the three-dimensional structure of NA. The biological significance of these residues is unknown. The two-step NA deletions—first 17 amino acids and then a further 7 amino acids—ocurred in concert with stepwise changes in HA, possibly to maintain the functional balance between HA and NA. Indeed, the first NA stalk deletion was associated with the introduction of N188D in HA, while the second was associated with the N159D mutation in HA.
The locations of amino acid changes around the receptor binding pocket of HA (Fig. 3B) suggested that receptor recognition may be altered. In addition, NA stalk deletions, which affect NA receptor-destroying activity (5, 29), were likely introduced to compensate for changes in receptor recognition. We therefore compared receptor recognition in the parent virus and in viruses passaged 19 times in quail: non-plaque-purified P19T and plaque-purified clones (P19T clones 2 and 4; P19T clone 2 possessed the representative HA of the P19T viruses) (Table 1). P19T clones 2 and 4 differed by only 2 amino acids (underlined) in HA [P19T clone 2 HA, I3L, N54K, A144D, N159D, N188D, S227P, E325G, G4S(HA2); P19T clone 4 HA, I3L, N54K, A144D, N137S, N188D, S227P, E325G, G4S(HA2)]. The first receptor assay we employed relied on the inhibition of virus binding to fetuin [which possesses both α(2-3)- and α(2-6)-linked sialic acids] with sialylglycoconjugate polymers that possess either α(2-3)- or α(2-6)-linked sialic acid (Fig. 4A) (21). Despite the mutations near the HA receptor binding site, this assay did not reveal any appreciable differences between the parent and the passaged viruses in terms of receptor recognition. Failure to detect differences in receptor specificity may stem from reliance on a single glycan structure, i.e., α(2-3)- or α(2-6)-sialyloligosaccharide in this assay. Therefore, we further assessed receptor specificity by glycan microarray (Fig. 4B). The microarrays displayed 86 sialylated glycans and 9 asialoglycans printed on coated glass slides. Although no major differences in receptor recognition between the parent and quail-passaged viruses were detected, the binding of P19Tcl4-HA for α2-6 biantenary sialosides (46 to 48) was slightly higher than that of the parent virus HA in all three independent experiments.
To examine the effects of the NA stalk deletions on NA function, we compared the activities of the NAs (P19Tcl2-NA with the first and second deletion and P19Tcl4-NA with only the first deletion) by using an erythrocyte elution assay (Fig. 5). No appreciable difference in the elution from chicken erythrocytes was noted among the viruses. In contrast, the viruses with quail-adapted NA, WT(HA)-P19Tcl2(NA)-RG and WT(HA)-P19Tcl4(NA)-RG, were eluted more slowly from guinea pig erythrocytes than the WT(HA,NA)-RG virus. Chicken erythrocytes contain more Sia2-3Gal linkage than Sia2-6Gal, and this is reversed in guinea pig erythrocytes (16, 28). This finding is consistent with previous findings that NA stalk deletions negatively affect the ability of virions to hydrolyze sialic acid from membrane-bound substrates (5, 29). In addition, we tested the sialidase activities of these quail-passaged viruses in a standard assay that uses 4-methylumbelliferyl-N-acety-lneuraminic acid as a soluble substrate (34). No appreciable differences were observed between the parent and the quail-passaged viruses, indicating that the stalk deletion, as expected (7), and the two mutations found in the later isolates (Q46P and I215V) do not affect removal of sialic acid from a soluble substrate (data not shown).
We then evaluated the effects of the mutations in HA and NA on the replicative potential of the viruses. To eliminate the effects of mutations that may have been introduced in genes other than the HA and NA genes, we compared the growth kinetics of the reassortant viruses that possessed the HA and NA of P19Tcl2 [P19Tcl2(HA,NA)-RG], P19Tcl4 [P19Tcl4(HA,NA)-RG], and wild-type [WT(HA,NA)-RG] viruses and the remaining genes from A/whistling swan/Shimane/499/83 (H5N3). In DEF, CEF, and MDCK cells (Fig. 6a, b, and c), the P19Tcl2(HA,NA)-RG virus grew to a higher titer than the WT(HA,NA)-RG and P19Tcl4(HA,NA)-RG viruses (Fig. 6 a, b, and c). The P19Tcl4(HA,NA)-RG virus also showed slightly higher replicative ability in CEF, but not in DEF and MDCK cells. This result suggests that the quail-adaptive mutations in the HA and NA genes of the virus may enhance its replicative ability in chicken cells. In human primary respiratory (NHBE) cells, all of these reassortant viruses grew too poorly to compare their replicative abilities. Therefore, we generated viruses possessing the internal genes of seasonal human H3N2 (A/Yokohama/2017/03) virus instead of those of the swan virus. Both P19Tcl2(HA,NA)-human-RG and P19Tcl4(HA,NA)-human-RG showed much higher replicative ability than the WT(HA,NA)-human-RG virus in NHBE cells (Fig. 6d). This result suggests that the changes that occur in the HA and NA of duck viruses during their adaptation to quail may allow the duck influenza viruses to replicate efficiently in human cells.
To determine whether the virus adapted to quail acquired the ability to efficiently infect human cells ex vivo, the parent and non-plaque-purified P19T viruses, as well as a control duck virus [A/duck/Vietnam/5001/04 (H5N1), which preferentially binds to α(2-3)-linked sialic acid (36)] and a human seasonal virus [A/Kawasaki/173/01 (H1N1), which preferentially binds to α(2-6)-linked sialic acid (36)] were used to infect human respiratory tissue (alveoli and bronchus). Although most of the human respiratory tract, including the bronchus, contains Siaα2-6Gal, Siaα2-3Gal is also present in human alveoli (32, 36, 45). Indeed, we detected expressed viral antigens in the alveolar cells incubated with all of the viruses tested (Fig. 7c, d, g, and h). However, in the bronchus, where only Siaα2-6Gal is present, expressed viral antigens were detected only when incubated with P19T (and the control seasonal human virus) and not in tissues incubated with the parent virus or the control H5N1 duck virus (Fig. 7a, b, e, and f). This result suggests that adaptation of duck viruses to quail may expand their host range not only to chickens (14, 22, 33, 38, 39), but also to humans.
Our initial studies were aimed at validating the experimental quail animal model and confirming that the epithelial cells lining most of the respiratory tract of quail possess mainly Siaα2-6Gal molecules, except for the epithelial cells in the parabronchus (9, 18, 37) (Fig. 1). The presence of both Siaα2-3Gal and Siaα2-6Gal receptors on the epithelial cells of the quail respiratory tract may explain why some avian H9N2 viruses isolated from quail recognize Siaα2-6Gal (26, 35, 43). To demonstrate the susceptibility of quail to infection with viruses from aquatic birds and the subsequent selection of variants with expanded host ranges under experimental conditions, we performed serial infections of quail by the respiratory route using a duck influenza virus subtype H3N2. During this process, the virus acquired the ability to grow well in its new host, particularly in respiratory organs. This preferential replication of waterfowl influenza virus in the respiratory tract of quail confirms a previous report (25). None of the infected quail showed clinical signs of disease, even though high virus titers were detected in respiratory organs. However, the latter would indicate that the adapted virus could be transmitted among quail without noticeable clinical signs.
During sequential passages in quail, the virus accumulated eight amino acid substitutions in the HA protein. Interestingly, the same substitutions have been found in the HA of human H3N2 viruses compared to their ancestral duck viruses. The human H3 HA acquired the N188D mutation in 1971, and it has been maintained since then. Similarly, the S227P mutation was detected in 2003 and has also been maintained ever since. The A144D mutation occurred in 1971 and was maintained until 1981. The N137S mutation occurred in 1999 and has been fixed since then. Among the eight mutations, 2 amino acids (at positions 137 and 227) are located near the receptor binding site. We previously showed that an amino acid substitution at position 223 (227 in the H3 numbering) and at position 134 (138 in the H3 numbering) allows avian H5 viruses to bind to human-type receptors (44). Thus, substitutions at these positions may contribute to a change in receptor recognition. The amino acid at position 159 is located at the very top of the HA globular head. A previous study showed that during adaptation to cultured human cells, some avian viruses acquire a negative charge in the top region of their HA, with substitutions by acidic amino acids, which results in a decreased affinity for membrane receptors (18). N159D and N188D may increase the local negative charge of the globular head region. Therefore, they may also contribute to adaptation to human-type receptors via this mechanism.
Although we did not detect significant differences in receptor specificity between the parent and quail-passaged viruses with the assays we employed, P19T clone 4 showed slightly higher binding affinity for some α2-6 biantenary sialosides (46 to 48) than the parent virus in our glycan microarray analysis. In addition, despite inferior replication in duck cells, the viruses that possessed the HA and NA of P19T clone 4 showed higher replicative abilities in NHBE cells than the virus with the original HA and NA. The slightly higher binding of P19Tcl4-HA for glycans 46 to 48 may contribute in part to the high growth of the virus that possessed the HA and NA of P19T clone 4 in NHBE cells. Further study is required to fully understand the adaptation of duck virus in quail. The virus possessing the HA and NA (P19T clone 2) of the virus passaged 19 times in quail also replicated efficiently in NHBE cells. More importantly, the quail-passaged viruses (but not the duck virus) infected the epithelial cells of the human bronchus, where only human-type receptors exist. These findings suggest that the mutations that occurred in HA during passage in quail likely contribute to the increased recognition of receptors that are expressed on human epithelial cells, although these changes in receptor recognition may be too subtle to be detected by the binding assays we employed.
During the passages in quail, stepwise NA stalk deletions were also introduced. Such deletions are often observed in avian viruses isolated from humans and terrestrial poultry, such as chickens and quail, but not aquatic birds (26, 38). The NA deletions also suggest a change in receptor binding affinity because NA stalk deletions were reported to reduce cleavage of membrane-bound sialosides by NA (29, 38). Our study also indicated that the NA stalk deletion impaired virus release from guinea pig RBCs, supporting the notion that NA adjusts its function to maintain balance with a changed HA. However, there was no appreciable difference between the parent and mutant viruses when chicken RBCs were used, suggesting that the effect of the NA stalk deletions on NA function may be dependent on the properties of receptors present on the host cells.
Amino acid changes at positions 325 (E325G) and 333 (G333S) were also introduced during passages in quail. These residues are located at positions −5 and +4 of the HA1-HA2 cleavage site. The Glu residues of the fusion peptide at positions 11 and 15 are thought to interact with the lipid membranes (8, 17, 22). The amino acid at position 333 (4 in HA2) is situated close to the amino acids at positions 11 and 15 in the helical wheel of the fusion peptide. These mutations may affect the fusion activity of HA and may have been introduced because the pH in the endosome may differ between quail and ducks. Alternatively, these mutations may be associated with the tissue tropism of the quail-passaged virus, which preferentially replicated in respiratory organs rather than the intestine; in this way, most of the viruses in quail are not exposed to the low pH in the stomach.
In this study, we could not detect appreciable shifts in receptor binding specificity during serial passage of the duck virus in quail. However, the HA and NA of quail-passaged viruses supported enhanced replication in human bronchus epithelial cells, but not duck cells, and quail-passaged virus acquired the ability to infect the human bronchus, where Siaα2-6Gal is the dominant receptor type. These findings strongly suggest that quail may be an intermediate host for aquatic-bird viruses to acquire the ability to replicate efficiently in chicken (14, 22, 33, 38, 39) and human cells.
We thank Susan Watson for editing the manuscript.
The glycan microarray was produced for the Centers for Disease Control and Prevention (CDC) by using a glycan library generously provided by the Consortium for Functional Glycomics funded by National Institute of General Medical Sciences Grant GM62116. This work was supported by a Grant-in-Aid for Specially Promoted Research; by a contract research fund for the Program of Founding Research Centers for Emerging and Reemerging Infectious Diseases from the Ministry of Education, Culture, Sports, Science, and Technology; by grants-in-aid from the Ministry of Health; by ERATO (Japan Science and Technology Agency); and by Precursory Research for Embryonic Science and Technology (PRESTO) and the Global Center of Excellence (G-COE) for Education and Research on Signal Transduction (Japan Science and Technology Agency); as well as by National Institute of Allergy and Infectious Diseases Public Health Service research grants.
The findings and conclusions in this report are those of the authors and do not necessarily represent the views of the Centers for Disease Control and Prevention or the Agency for Toxic Substances and Disease Registry.
Published ahead of print 16 November 2011
Supplemental material for this article may be found at http://jvi.asm.org.