Avian influenza A (H9N2): computational molecular analysis and phylogenetic characterization of viral surface proteins isolated between 1997 and 2009 from the human population

It has been well documented that the receptor binding site motif of HA is critical for cellular receptor specificity and determining virus host range [21, 22]. Out of five conserved amino acids in the pocket of the HA glycoprotein, two positions showed the maximum number of mutations (198, 234) and three remained 99% conserved (191, 235, and 236). The pattern of observed mutations at position 198 was E¹⁹⁸T, E¹⁹⁸A, E¹⁹⁸V and E¹⁹⁸D, whereas for position 234 was L²³⁴Q, Q²³⁴L and Q²³⁴M as summarized in Table 2. Mutations in these regions are considered strong factors for change in sugar specificity thus leading to change in host specificity. The presence of glutamine (Q) at position 234 (H3 numbering: 226) is a typical avian virus signature, and it has been reported that presence of this amino acid results in a preference for binding to 2,3-linked sialic acid (avian receptors) whereas, in the case of leucine (L) at the same position, there is a preference for 2,6-linked sialic acid (human receptors) and potential cause of reported human infections [5, 6, 23‐25]. However, upon analysis of amino acids at the receptor binding site of HA glycoproteins, we have identified that H9N2 viruses isolated in 2009 from infected patients in Hong Kong possessed glutamine at position 234 (H3 numbering: 226) instead of leucine as has been found in previous isolates, yet still somehow managed to bind with human sialic receptors. As shown in Table 2, this residue variation in avian H9N2 viruses has not been reported before in human cases. This observation can be correlated with a number of phenomena such as genetic evolution of influenza viruses in order to evade host defense mechanisms, and transfer of this virus from swine to humans instead of avian to human transmission. It is also known that pigs act as a "mixing vessel" because viruses isolated from pigs recognize both types of sialic receptors [26] and most importantly represent the balance that exists between the activities of HA and NA. Both the genes PB2 and HA are known to be critical for the pathogenicity of the virus. Therefore, the role of other proteins such as the PB2 segment [4, 27] must be viewed as potential causes of zoonotic H9N2 possibly resulting in future human to human transmission.

Analysis of the HA cleavage site showed that strains isolated in 2009 have a different cleavage site compare to those previously reported. From 1997 to 2008, all H9N2 viruses retained a conserved amino acid pattern at the cleavage site; ³³⁵RSSR³³⁸GLF³⁴¹ except for the sublineage prototype strain (A/chicken/Kr/38349-p96323/96) that had the cleavage site ³³⁵ASYR³³⁸GLF³⁴¹. Presence of the R-S-S-R motif is suggestive of H9N2 viruses with low pathogenicity that have adapted to chicken host [28‐30]. However, it has been observed that viruses isolated in 2009 have the different pattern (³³⁵RSNR³³⁸GLF³⁴¹) due to the substitution mutation S³³⁷N (Table 2). To our knowledge, this is the first report to show presence of the R-S-N-R motif in avian H9N2 influenza viruses isolated from human patients although the R-S-N-R motif had been reported before in different studies on avian H9N2 viruses from around the globe [31‐33]. The significance of this mutation on viral stability or increased pathogenicity is not fully understood as yet and requires further studies, as the cleavage site is considered an indicator of pathogenicity [34, 35]. Comparison with vaccines strains also revealed variations at positions 198 and 234 in the HA glycoprotein. The currently available vaccine virus to the virus has an E and an L at positions 198 and 234, respectively, whereas strains associated with human infections reported in 2009 contain D and Q residues at these respective positions.

The major factors known to influence the functional activities of the NA glycoprotein are enzyme active sites, the stalk length, HB site and potential glycosylation sites. The HB site is located on the surface of the NA molecule, away from the neuraminidase enzyme active site [36]. Analysis of the HB site showed substitutions similar to those detected in the Y280 lineage prototype strain (A/Duck/Hong Kong/Y280/97) and in H9N2 human isolates from Hong Kong in 1999. These mutations are typical of human pandemic H2N2 and H3N2 viruses [7, 14]. The most prominent mutation is at position 402 from isoleucine to asparagine/serine and at position 403 from arginine to tryptophan/serine. Strains from 1999 - 2009 also showed mutations at position 367 from lysine to glutamic acid and at position 432 from glutamine to lysine as described in Table 2. However, the biological significance of any of these substitutions in the HB site is not yet known.

The NA protein consists of a box-like head with an enzymatic active site that is connected to a fibrous stalk region of variable length. The NA stalk is important for balancing the complementary activity of HA and NA and has been correlated with efficiency of virus replication and pathogenesis. From various studies previously performed it has been concluded that longer stalk length of the virus results in better replication [37‐40]. Analysis of stalk length revealed that the two prototype viruses, Qa/Hong Kong/G1/97 and A/Hong Kong/1073/97, of the G1-lineage contained a two amino acid deletion in the NA stalk region at positions 38 and 39, a deletion at position 62 - 64 was observed in A/Duck/Hong Kong/Y280/97, A/Shaoguan/408/98 and A/Shaoguan/447/98. The particular 46-50 amino acid deletion, which is important for poultry adaptation of the virus [41], is not found in any of the analyzed sequences even though these viruses are thought to be transmitted from avian species to humans. Furthermore, as shown in Table 2, from 1999 - 2009 none of the reported strains contained a stalk deletion at any position, which may also be an indication of evolution in the viral structure leading to a NA protein with better replication rate in humans.

Sequence analysis of binding pocket residues in NA for drugs such as zanamivir (Relenza^®) and oseltamivir (Tamiflu^®) was also performed. These analyses showed that for each virus, amino acids in the enzyme active site were conserved and no substitution mutations were seen, which may lead to an evolution of sialidase inhibitor resistant viruses (Table 2). It is known that mutations at positions other than the active sites still tend to alter the activity of active site amino acids; therefore this possibility cannot be excluded. Such mutations are not always captured or extensively studied in homology models. Therefore, wet lab studies are required to check the level of effectiveness possessed by NA inhibitors against recently reported strains.