Hemagglutinin (HA) is a protein found in the membrane of a virus that mediates receptor binding and entry of the genome-transcriptase complex into cells during infection (1).  With the recent infections of avian influenza in Homo sapiens in China in 2013, the binding affinity of receptor analogues with human HA (PDB ID: 4BSA) is an important topic of research.  In accordance with many studies, small residue differences between HAs of different species causes new conformations between the receptor and membrane protein and changes in binding avidity between human and avian H7 influenza.  By analyzing the binding preferences between both human and avian HAs with human and avian influenzas (H7N9 and H7N3, respectively), scientists hope to understand why a pandemic did not occur in China and if one could occur in the future (2). 

Hemagglutinin is a protein consisting of alpha helices, beta strands, and random coils. This secondary structure folds into subunits that are 13.5 nanometers long, which are split into two domains, which include acid, basic, polar, and hydrophobic regions (3).  Folded into a globular conformation, the membrane-distal domain is used by the influenza virus to bind to its host’s cells and is encoded by residues 68-195 of the A subunit.  Each of the distal domains contains a site that binds sialic acid on the host target molecules of airway epithelial cells in humans.  HAs of different species prefer different sialic acid to galactose linkage.  For example, avian influenza viruses bind preferentially to alpha(2,3)-linked sialic acid, while human viruses recognize alpha(2,6)-linked sialic acid (see Figure 1).  A study found that Leu-226 is responsible for recognizing alpha(2,6) sialic acid, demonstrating the just one nucleotide can decide the specificity of an HA (4). The receptor binding portion of the HA protein is encoded by residues 190-198 to form the 190-helix;residues 133-138 to form the 130-loop; and residues 220-229 the form the 220-loop.  Residues Tyr-98, Trp-153, His-183, and Tyr-195 are conserved to form the basis of the binding site. (5). The preference in humans is because the alpha(2,6)-linked sialic acid  (see Figure 1) is the major form found in the human respiratory tract (1). The membrane-proximal domain of HA, adjacent to the viral membrane, consists of an alignment of two long alpha helices with beta strands. Long loops and short turns join the helices and strands at the surface of the molecule.  The quaternary structure consists of three subunits that are stabilized by lateral hydrogen bonding interactions, forming a triple-stranded coiled-coil.  In addition to hydrogen bonding, HA also contains many covalent bonds in carbohydrate chains, characteristic of membrane proteins (3). 

Two different systems are available to find comparison proteins to human hemagglutinin.  PSI-Blast compares the primary structures of proteins by analyzing the sequence homology and assigning an E value to the comparison protein.  A gap in the sequence is caused by an amino acid or group of amino acids that is present in the subject sequence but not in the search sequence.  This causes an increase in the E value.  On the other hand, homology decreases the value.  Any E value less than 0.05 is significant for proteins, and many proteins compared to human hemagglutinin, have an E value of 0.0.  Although not crystallized, other HAs associated with various avian influenzas have an E value of 0 because the primary sequence only differs by one or a few nucleotides (7).  The Dali server, on the other hand, compares the tertiary structure of a subject protein to other proteins by analyzing intramolecular distances and assigning a Z value.  Any Z value over two is considered substantial. Dali matched the primary protein of human HA in complex with the avian receptor to the avian HA in complex with the human receptor (PDB ID: 4BSH) with a Z score of 41.5 (8).  The small difference between the Z scores demonstrates the minimal modification of the tertiary structure of the two binding pockets. The results of these scores with comparison proteins show that the HA H7 complex is very similar between different species, and this different is due to very small changes in the amino acid sequence. 

These small differences in the primary sequence of the hemagglutinin protein can cause large changes in the binding avidity and conformation of the HA and sialic acid receptor analogue bond.  For example, in H13 avian influenza, a single nucleotide change at position 186 from polar, uncharged asparagine to hydrophobic valine causes the virus to specifically bind to the avian receptor and not the human receptor.  When asparagine is in that position, the binding to the avian receptor decreases and the binding to the human receptor increases (6).  Another study found that for H2 and H3 viruses, mutations of Glu-226 and Gly-228 to leucine and serine, respectively, causes a shift from avian to human receptor specificity (5).  For the H7 virus, the glutamine residue at position 226, characteristic of avian viruses, is replaced by hydrophobic leucine. In addition, the glycine residue at position 186 in the avian HA changes to a hydrophobic, bulkier valine in the human HA. These two changes cause a specific part of the binding site where galactose-2 sits to be more hydrophobic (2). These examples show that even single nucleotide differences can change the binding conformation of avian and human sialic acids to avian and human HAs. 

In the binding pocket of HA, the H7 influenza virus is binding sialic acid, galactose-2 and partial density of a third sugar, N-acetyl glucosamine.  The small primary sequence changes between avian and human receptor with H7 change the conformation of binding with these three sugars, changing the avidity of the virus and success of infection.  When the human receptor binds with the human H7 HA, good electron density allows the receptor complex to form, and the receptor projects out of the binding site.  When the human receptor binds with avian H7 HA, the electron density is weaker for sialic acid and galactose, causing the galactose to appear face-on instead of edge-on in the complex.  This causes a forty-five degree difference around the galactose-2 C6-C5 bond.  All of these changes are caused by the substitution of the bulky leucine and valine amino acids, which make the binding pocket substantially more hydrophobic.  This increased hydrophobicity widens the gap under the C6 and galactose-2 region between the 220-loop and 130-loop by about one angstrom. The valine substitutionprecludes two water molecules, which makes galactose-2 sit further out in the binding pocket.  When the sialic acid and other sugars of the avian cell bind human H7 HA, it is in cis, but when it binds avian H7 HA, it flips to trans. This change is also due to the hydrophobic bulky leucine introduced by the human HA (1). 

The introduction of the bulky, hydrophobic residuesLeu-226 and Val-186 are responsible for H7N9 not becoming a pandemic in the human population.  This residue difference between the human HA and avian HA changes the affinity for sialic acid linkages, but not enough to cause a pandemic.  In order to transfer between mammals, the H7 influenza virus must exponentially prefer human host cells to avian cells, which is not the case.  While H7 has a higher affinity for the alpha-2,6-linked sialic acid found in human respiratory cells, it still retains a strong affinity for the avian alpha-2,3-linked sialic acid of the avian receptor. (see Figure 1) The “avian” receptor sialic acid linkage is present in mucins in the respiratory tract of humans, so the preference of the H7 virus for avian receptors restricts binding and limits the infection.  In order to become pandemic, the H7 virus would have to completely lose its affinity for avian sialic acid linkages while retained its strong affinity for the sialic acid linkages in humans.  This was the case with the aerosol-transmitted H5 virus because it completely lost avidity for the avian receptor.  In addition, there is a difference in pH needed to cause conformation changes in HA to occur between the HAs of H7 and those of a pandemic virus.  The pH is higher for human H7 HA than that of the HAs of pandemic virus, 5.6 versus 6.3, respectively.  This reveals that the human H7 HA is less thermostable, so a mutation would be required to increase its stability before it could be passed from human to human.  This mutation was also achieved with the H5 virus by a mutation at Thr318Ile, which increased thermostability and allowed it to become pandemic and pass between humans (1). 

Theminute primary sequence differences between the membrane protein hemagglutinin of human H7N9 influenza and avian H7N3 influenza cause conformational changes in the binding pocket that changes the binding affinities with the sialic acid of receptor cells.  The introduction of two bulky, hydrophobic bases, leucine and valine,change the binding pocket complex enough to cause the human HA to prefer the human receptor, alpha(2,6)-linked sialic acid (see Figure 1). This alteration is not substantial enough to eliminate the strong binding between human HA and the avian receptor, alpha(2,6)-linked sialic acid, which hinders the H7N9 influenza from becoming transferable between mammals and reaching a pandemic level.