Bioinformatics and Structural Analysis of Antigenic Variation in the Hemagglutinin Gene of the Influenza A(H1N1)pdm09 Virus Circulating in Shiraz (2013 to 2015)

ABSTRACT Circulating influenza A virus provided an excellent opportunity to study the adaptation of the influenza A(H1N1)pdm09 virus to the human host. Particularly, due to the availability of sequences taken from isolates, we could monitor amino acid changes and the stability of mutations that occurred in hemagglutinin (HA). HA is crucial to viral infection because it binds to ciliated cell receptors and mediates the fusion of cells and viral membranes; because antibodies that bind to HA may block virus entry to the cell, this protein is subjected to high selective pressure. In this study, the locations of mutations in the structures of mutant HA were analyzed and the three-dimensional (3D) structures of these mutations were modeled in I-TASSER. Also, the location of these mutations was visualized and studied using Swiss PDB Viewer software and the PyMOL Molecular Graphics System. The crystal structure of the HA from A/California/07/2009 (3LZG) was used for further analysis. The new noncovalent bond formations in mutant luciferases were analyzed via WHAT IF and PIC, and protein stability was evaluated in the iStable server. We identified 33 and 23 mutations in A/Shiraz/106/2015 and A/California/07/2009 isolates, respectively; some mutations are located on the antigenic sites of Sa, Sb, Ca1, Ca2, and Cb HA1 and the fusion peptide of HA2. The results show that with the mutation some interactions are lost and new interactions are formed with other amino acids. The results of the free-energy analysis suggested that these new interactions have a destabilizing effect, which needs confirmation experimentally. IMPORTANCE Due to the fact that the mutations that occurred in the influenza virus HA cause the instability of the protein produced by the virus and antigenic changes and the escape of the virus from the immune system, the mutations that occurred in A/Shiraz/1/2013 were investigated in terms of energy level and stability. The mutations located in a globular portion of the HA are S188T, Q191H, S270P, K285Q, and P299L. On the other hand, the E374K, E46K-B, S124N-B, and I321V mutations are located in the stem portion of the HA (HA2). The change V252L mutation eliminates interactions with Ala181, Phe147, Leu151, and Trp153 and forms new interactions with Gly195, Asn264, Phe161, Met244, Tyr246, Leu165, and Trp167 which can change the stability of the HA structure. The K166Q mutation, which is located within the antigenic site Sa, causes the virus to escape from the immune response.

generally produce similar symptoms, but their antigenic symptoms are not related to each other. Hence, infection with one type confers no immunity against the others (3). The influenza type A viruses cause great influenza epidemics, the influenza type B viruses cause smaller localized outbreaks, the influenza type C viruses cause only mild respiratory illness in humans, and the influenza type D viruses are not known to infect humans (4). Influenza A viruses are classified into subtypes that are differentiated mainly on the basis of two surface antigens, hemagglutinin (HA) and neuraminidase (NA). These proteins mediate host cell attachment and release. HA attaches virions to cells by binding to terminal sialic acid residues on glycoproteins/glycolipids to initiate the infectious cycle, while NA cleaves terminal sialic acids, releasing virions to complete the infectious cycle. These proteins are the primary targets of the protective antibody-mediated immune response (5). HA has functionally defined immunodominant antigenic sites that primarily map to the globular domain of the glycoprotein and surround the receptor binding site (RBS) (6). Circulation of influenza viruses gradually aggregates HA mutations (antigenic drift) or segment swapping (antigenic shift) in the antigenic sites targeted by neutralizing antibodies, allowing these mutations to escape the immune response and vaccination (5). Once that influenza virus acquires an HA protein through reconstruction, influenza pandemics may occur (7).
The influenza A(H1N1)pdm09 virus has been evolving since April 2009, acquiring new amino acid changes that may alter its antiviral drug susceptibility and antigenic and virulence characteristics (8). The protein structure of HA is composed of three monomers, each monomer composed of a heavy (HA1) (;40-kDa) and a light (HA2) (;20-kDa) chain that is held together by a disulfide bridge and noncovalent interactions (9). HA1 forms the receptor binding sites and antigenic sites (Sa, Sb, Ca1, Ca2, and Cb), while HA2 contains a conserved fusion peptide (10). These antigenic site residues are highly variable, and mutations are tolerated by viruses (11). Phylogenetic analysis of the HA gene of influenza A(H1N1)pdm09 viruses showed that it clustered into 8 genetic groups (12). Mutations in influenza virus HA proteins may alter the activity of influenza vaccines and antiviral drugs and are one of the possible catalysts for previous world pandemics (5). Many antiviral drug products, including vaccines, monoclonal antibodies (MAbs), and NA inhibitors (NAIs), target the HA or NA glycoproteins (5). Due to changes in the dynamic HA, the activity of these vaccines and drugs may be affected and the viral NA or HA protein mutations have reduced susceptibility to NAIs. For example, the D222G mutation in the HA1 subunit of HA was associated with severe clinical outcomes (12). These mutated viruses may be less transmissible but more pneumotropic and more resistant to antiviral treatment (13,14). This resistance was generally associated with other mutations in the NA protein, although other mutations were also described to confer resistance to NAIs (15). Global virology monitoring needs to be updated for influenza virus mutations that may affect viral characteristics such as virulence, transmission, or antiviral susceptibility. NAI mutations in HA are likely to have the effect of lowering the receptor binding avidity and compensate for decreased activity of NA (16). However, it is not clear whether HA mutations are associated with decreased immune reactivity to anti-influenza virus antibodies.
Since the discovery of the first case of influenza A(H1N1)pdm09 in Fars province, Iran (15 July 2009), there have been many reports of influenza A(H1N1)pdm09 in patients with suspected influenza (14). As the HA mutations may affect the receptor binding specificity and strain pathogenicity, for the identification of emerging variants, continued epidemiological and molecular studies are essential for monitoring the modifications in the virus genome (17,18). Also, evaluating the structural changes caused by these mutations in the protein is very important in understanding the danger level of the mutations and how the pathogenicity changes in this virus. Furthermore, the effects of these mutations on the function of the HA molecule were not studied. In this regard, we previously conducted a molecular and phylogenetic analysis of new influenza A/H1N1 (A/Shiraz/2013 and A/Shiraz/2015) virus strains that circulated in Fars province (19). In the present study, we investigated these identified HA mutations that may lead to changes in antigenic profiles and affect antibody-mediated virus inhibition. Amino acid sequence analysis of the HA gene (amino acids 1 to 566) indicated mutations in the HA1 domain, including Sa, Sb, Ca, and Cb (20), and in the HA2 region. The HA genome shows the highest mutation rates that result in influenza virus genome instability (21). In this regard, the purpose of this study was to analyze the location of these identified mutations in the structure of HA. Furthermore, the three-dimensional (3D) structure of these mutated HAs was modeled in I-TASSER (20) and the location of these mutations was studied using Swiss PDB Viewer software (22) and the PyMOL Molecular Graphics System (23). The crystal structure of the HA from A/California/07/2009 (3LZG) (24) was used for further analysis. The new noncovalent bond formations in mutant luciferases were analyzed via WHAT IF (25) and PIC (26). In the following experiments, the effect of these mutations on protein stability was evaluated in the iStable server. These results help in elucidating the HA folding mechanism, as well as the rational design of new and effective drugs and diagnostic reagents. These data emphasize monitoring the process of influenza virus changes so that the vaccine composition can be changed according to the circulating strains.

RESULTS AND DISCUSSION
MSA. The amino acid sequence reference and mutated HA (ARI70442.1 and AIE52254.1, respectively) were retrieved and aligned in the Clustal Omega tool (https://www.ebi.ac.uk/ Tools/msa/clustalo/). The multiple sequence alignment (MSA) was saved in the clustal_num format and analyzed with the Jalview software ( Fig. 1 and data not shown). The crystal structure of the HA region from A/California/07/2009 (3LZG) was used for further analysis of the sequence alignment. The results show that some mutations compared to the A/ California/07/2009 sample have been repeated in both the A/Shiraz/1/2013 (AIE52254.1) and A/Shiraz/106/2015 sequences. However, some mutations have occurred in only one of these. Also, the sequence of A/Shiraz/1/2013 (AIE52254.1) and A/Shiraz/106/2015 is about 45 nucleotides longer than the A/California/07/2009 sample ( Fig. 1).
Study of mutation effect on protein stability. A relatively high number of mutations have been identified in the sequenced influenza A(H1N1)pdm09 virus that was collected from 2013 to April 2015 in Shiraz, Iran. The mutations are located in the HA1 (the globular portion of HA) and HA2 (stem portion of HA) domains of HA. The most common mutations were identified at the antigenic sites Ca, Sb, and Sa. Previous studies show that a single mutation by change of free energy (DG) may change the protein structural stability, while the difference in DDG between wild-type and mutant proteins is an impact factor in protein stability changes (27). In this regard, the effects of these mutations were evaluated in the iStable server. Free-energy analysis showed that most of these mutations have a slightly destabilizing effect (Table 1). It should be noted that in the iStable server, from the five webbased prediction tools that were chosen as element predictors, the final free-energy change was reported from the Meta result. While some of these element predictors may show positive results, the general conclusion reported is relevant to the Meta result that shows a single mutation by change of free energy may impact the protein structure stability. Some of these analyzed mutations are located in the stalk region. For universal influenza virus vaccines, the highly conserved HA stalk domain is an attractive target (28). In animal models antistalk antibodies are known to protect against a wide range of influenza viruses (29).
Bioinformatics studies. For a better analysis of these mutations, the 3D structures of this HA were modeled in the I-TASSER server; the best model was selected (Fig. 2). These   Table 2 shows the physicochemical properties of HA from A/ Shiraz/1/2013 that were calculated by the ProtParam tool. The results of this table show that the abundances of positively and negatively charged amino acids in the two proteins are similar. Instability index is a measure of proteins, used to determine the stability in a test tube. If the index is less than 40, then the protein is probably stable in the test tube. According to Table 2 the hemagglutinin of A/Shiraz/1/2013 is more stable than the hemagglutinin of A/Shiraz/106/2015. Figures 2B and C show the three-dimensional models of HA from the influenza A/ Shiraz/1/2013 H1N1 and A/Shiraz/106/2015 H1N1 viruses. These models were superimposed on the crystal structure of the 2009 H1N1 pandemic influenza virus (PDB: 3LZG) ( Fig. 2A). The 3D structure analyses showed that the structures of these models have a   (Fig. 2D).

Hemagglutinin Antigenic Variation in H1N1
Microbiology Spectrum mutation is located in the globular part of HA and may have a role in the structural stability of HA (10) (Fig. 4). Mutation of Q191H may eliminate interactions with T200, N250, T187, S188, L194, Y195, and Q196 residues and forms a smaller number of interactions with His205, Ser207, Leu208, and Tyr209, which can reduce the stability of the HA structure (Table 3). Areas outside the defined antigenic sites on the HA head can also be binding sites for antibodies (30). In many cases, these outside regions are more conserved and antibodies targeting these regions may provide more protection against influenza virus (31). V252 and Leu252 are located outside the defined antigenic sites, have a similar functional group, and do not seem to cause much structural change (Fig. 5). Freeenergy analysis showed that the V252L mutation has a structural destabilizing effect ( Table 1). The change V252L mutation eliminates interactions with A181, F147, L151, and Trp153 and forms new interactions with G195, N264, F161, M244, Y246, L165, and W167, which change the stability of the HA structure ( Fig. 5 and Table 4).
Mutation K285Q is located in a basic patch in the stalk region of HA (hemagglutinin A/Shiraz/1/2013). Due to the mutation, K285's interactions with the backbone of I297, T283, and H47 are lost. However, Q285 can interact with H313, T298, and H54 ( Fig. 6 and Table 5). This has a destabilizing effect (Table 5).
This study is a continuation of the previous research that analyzed the molecular diversity of pandemic influenza A(H1N1) viruses in Shiraz (19). As previous studies have shown, the surface proteins of influenza viruses must avoid recognition by the host immune system by modulating their activity and changing the positions of their surface glycoproteins (10). One  of the main mechanisms of modulating is antigenic drift, and the influenza A(H1N1)pdm09 virus is an effective model because of the extensive data available for the virus. In this study, we analyzed the structural changes caused by amino acid mutations in the HA of the prevalent influenza A(H1N1)pdm09 virus, which might help the course of influenza treatment.
In the previous study (19), we identified a number of mutations, some of which were previously reported and resulted in antigenic drift. These mutations are located on the antigenic sites Sa, Sb, Ca1, Ca2, and Cb in HA1 and the fusion peptide of HA2. As previously shown, the K166Q mutation is located within the antigenic site Sa (10). This mutation allows the virus to escape from the immune response. The S188T mutation is located in antigenic site Sb and has little impact on the antigenic drift. However, the accumulation of several mutations in different antigenic sites can result in a reduction in vaccine effectiveness.
Previous studies show that in the influenza A(H1N1)pdm09 virus the HA region is unstable (32). This instability leads to various effects on the structure of HA that may benefit the virus (33). In this regard, the mutations, by introducing or disrupting interactions within and between monomers, may have different effects on the stability of the structure of HA. Single point mutations or recruitment of several mutations has different impacts on HA stability.
In the D104N and D472N mutations, deletion of an acidic amino acid masks the acidic patch located in this region. With mutations of K123N, K165Q, and K285Q, the basic patch of this region is obscured. In the N444D and H140R mutations, new acidic and basic patches are formed. In E374K and E499K replacement of an acidic amino acid with an essential one masks the basic patch. Our results show that the alteration of residue interactions at mutation sites causes destabilization of the HA protein. However, this still needs experimental confirmation. An important aspect of influenza virus infection of airway epithelial cells is the interaction between the virus hemagglutinin (HA) protein and the corresponding receptor on the host cell. Although the precise nature of the viral receptor is incompletely defined, influenza viruses target glycosylated oligosaccharides that terminate in a sialic acid (SA) residue. These residues are bound to glycans through a2,3, a2,6, or a2,8 linkage by sialyltransferases that are expressed in a cell-and species-specific manner. Influenza viruses primarily target airway epithelial cells via a2,3and a2,6-type receptors, but the distribution of these receptors in many species is uncertain and may be a significant factor influencing infection. For example, influenza A viruses isolated from avian species  preferentially bind to SA receptors that are linked to galactose by Neu5Aca2,3Gal residues, while human strains preferentially bind the Neu5Aca2,6Gal-terminated sugar chains. Antibodies are one of the most essential types of therapeutic methods. Predicting the antibody-antigen interaction changes upon mutations (DDG binding) is essential for antibody engineering. On the other hand, identifying hot spots is crucial in the antibody-antigen interaction, because as the hot spots are mutated, the affinity is significantly reduced (34). Kinetic, thermodynamic, and structural analyses of mutants provide essential information about the individual amino acids and elucidate the characteristics of the respective antibody-antigen complexes (35). In the phase of vaccine engineering, obtaining structural, thermodynamic, and kinetic data from mutations is crucial. According to the bioinformatics analyses, the interactions of the mutated amino acids with other amino acids in the structure of HA have changed. This change of interactions can lead to a change in the position of the amino acids that affect the binding of the vaccine to the virus. These structural changes of HA have progressed toward instability based on the bioinformatics analysis of the structure. However, these analyses indicate that current influenza vaccines are likely to have a decreased efficacy in the future. The results of this study can be considered for the effective control of pandemic influenza and the development of a more appropriate vaccine. This study emphasizes the importance of bioinformatics analysis in the designing of drugs or vaccines before production.

Hemagglutinin Antigenic Variation in H1N1
Microbiology Spectrum their multiple sequence alignment (MSA) was performed by the ClustalW2 program (https://www.ebi.ac.uk/ Tools/msa/clustalo/) and MEGA7 (36). Molecular modeling of mutated HA. For in silico studies, the crystal structure of the wild-type HA was obtained from the Protein Data Bank (PDB). Also, the 3D structures of mutated HAs were modeled in the I-TASSER web server (20). On this web server, the 3D models are made based on multiple-threading alignments by LOMETS (37), and we chose the model with the best overall confidence score and Z-score from I-TASSER. The locations of mutations were visualized and studied using the PyMOL Molecular Graphics System (23) and the Swiss PDB Viewer (38). The ExPASy ProtParam server (39) was used to compute the physicochemical parameters of these models such as the molecular weight, the aliphatic index, the total number of positive and negative residues, the extinction coefficient, the theoretical isoelectric point (pI), the instability index, and the grand average hydropathy (GRAVY). The root mean square deviation (RMSD) between the corresponding atoms of two proteins was used to measure the similarity between two protein structures. The template modeling score (TM-score) is a measure of similarity between two protein structures. Hydrogen bonds were also calculated by the WHAT IF web server (25) and the PIC web server (26).
Study of mutation effect on protein stability. The effect of these mutations on the stability of mutated HA was evaluated on the iStable server (http://predictor.nchu.edu.tw/iStable/). The iStable server is a web server that predicts the value of free-energy change (DDG) using a support vector machine as an integrator algorithm (40). The iStable server integrates the results of i-Mutant2.0, AUTO-MUTE, MUpro, PoPMuSiC, and CUPSAT programs in the stability prediction and, with the support of a support vector machine, evaluates the point mutation effects on protein stability.