The use of pyrosequencing for detection of hemagglutinin mutations associated with increased pathogenicity of H5N1 avian influenza viruses in mammals

Abstract

Hemagglutinin (HA) cleavage is critical for virulence of influenza viruses. The amino acid residue at the P6 position of the HA cleavage site (HACS) has been shown to be most variable and to have a direct correlation with the cleavage efficiency and pathogenicity of H5N1 avian influenza viruses (AIVs) in mammals. Among these amino acid variants, serine has been associated with the highest virulence in mammals, and its detection may serve as an indicator for H5N1 AIVs with high pathogenicity and potential public risk. We developed a rapid detection method based on reverse-transcription (RT)-PCR and pyrosequencing to detect a mutation at the HACS that is associated with increased pathogenicity of H5N1 AIVs in mammals. Herein, we provide a specific, sensitive, and reliable method for rapid detection of one of the virulence determinants associated with increased pathogenicity of H5N1 AIVs in mammals.

Highly pathogenic H5N1 avian influenza viruses (AIVs; species Influenza A virus) have spread widely among domestic poultry in Asia, Europe, and Africa, causing severe mortality in avian species and sporadic fatal infections in humans.^9,13 As of 16 January 2017, highly pathogenic H5N1 influenza viruses have infected 856 humans, and 452 of these cases were fatal according to the World Health Organization (https://goo.gl/emYyde), raising public health concerns. Although few cases of human-to-human transmission have been reported, the highly pathogenic H5N1 AIVs may have potential to cause future influenza pandemics.

Hemagglutinin (HA) protein is responsible for influenza virus attachment on the cell surface.² Additionally, HA must be cleaved into HA1 and HA2 subunits to mediate membrane fusion between the virion and endosome.^16,17 The HA cleavage site (HACS) position was defined with cleavage occurring between P1 and P1’. The position numbers increase in the N-terminal direction along the cleaved peptide bond (P2, P3, P4, P5, and P6).¹¹ HA cleavage is critical for virulence of influenza viruses. HA proteins of highly pathogenic influenza viruses possess multiple basic amino acids at the HACS, which can be recognized by a ubiquitous protease. This cleavage permits systemic infection causing high mortality in poultry.⁷ In a previous analysis of the H5N1 HACS, we demonstrated that the amino acid residue at the P6 position of the cleavage site was the most varied, containing glycine (G), isoleucine (I), arginine (R), serine (S), or a deletion (*).¹⁸ Furthermore, the amino acid residue at the P6 position had a direct correlation with the cleavage efficiency and pathogenicity of H5N1 AIVs in mice. It is worth noting that serine is a typical P6 position amino acid residue of clade 2.1.3 H5N1 viruses that are predominant in Indonesia, which is the country with the highest case number and mortality rate of H5N1 viruses. Almost all Indonesian human isolates belong to clade 2.1.3, demonstrating that serine at the P6 position of the HACS may contribute to the virulence of clade 2.1.3 H5N1 in Indonesia.¹⁸ Therefore, serine at the P6 position of the HACS may serve as an indicator for H5N1 AIVs with increased pathogenicity in mammals and potential public risk.

Pyrosequencing is a new method of sequence identification and virus typing, which has been used for single nucleotide polymorphism (SNP) detection, screening antiviral-resistant mutations, and virus typing.^4,8,10 In a comparison with Sanger sequencing, pyrosequencing is more suitable for detecting sequences of dozens of bases when longer sequences are not required. To develop a pyrosequencing method to detect the HA virulence determinants associated with increased pathogenicity of H5N1 AIVs in mammals, we used multiple-sequence alignment to analyze the 1,303 HA gene sequences of H5N1 AIV listed in the Influenza Virus Resource of World Health Organization (https://goo.gl/71exJU). The alignment results indicated that the amino acid residue at the P6 position varied. For clade 2.1.3 viruses, 79.2% had serine at the P6 position; 78.6% of clade 7 viruses had glycine; and 95.7% of clade 5 viruses had isoleucine. Clades 0 (95.7%), 1 (84.0%), 2.1.1 (92.0%), 2.1.2 (100%), 2.2 (97.1%), 2.4 (92.0%), 3 (96.7%), 4 (100%), 6 (100%), 8 (100%), and 9 (100%) had arginine. Most of clade 2.3 (98.8–100%) and 2.5 (71.4%) viruses had one amino acid deletion at position P6. We also found that the adjacent amino acid residues of the HACS among different clades are conserved; only the residue at the P6 position is variable. Therefore, a set of RT-PCR primers specific for the HA consensus sequence, including the cleavage position, were designed (Primer 5.0, Premier Biosoft International, Palo Alto, CA). These primers amplify a 170-bp HA gene fragment that can identify different H5 clades (Table 1). The forward PCR primer was biotinylated in a manner suitable for the pyrosequencing assay. Next, a sequencing primer was designed (Pyrosequencing Assay (PSQ) Design v.1.0.6, Qiagen, Valencia, CA) to identify a 50-bp region covering the cleavage site (Table 1).

Table 1.

RT-PCR and sequencing primers used in the current study.

Primer	Sequence (5′–3′)	Target fragment size
Forward primer*	TCCACAACATACACCCTCTCAC	170 bp
Reverse primer	TACCATTCCCTgCCATCCTC
Sequencing primer	CCTgCTATAgCTCCAAA	50 bp

^*This primer is biotinylated.

To detect the various amino acid residues at the P6 position, we established a reverse genetics system for avirulent clade 7 H5N1 virus A/chicken/Sheny/0606/2008 (SY08). Briefly, all 8 gene segments were amplified by RT-PCR from SY08 and cloned into dual-promoter plasmid PHW2000. Depending on the polymorphism at the P6 position of the HACS, mutations (isoleucine, arginine, serine, or a deletion) were introduced into the HACS using a site-directed QuikChange mutagenesis kit (Agilent, Santa Clara, CA) according to the manufacturer’s instructions.¹⁸ The reassortant viruses were rescued by reverse genetics and designated rg325G (wild type), rg325I, rg325R, rg325S, and rg325*.

Viral RNA extraction and RT reactions were performed as described previously.¹⁵ PCR reaction was carried out following the manufacturer’s instructions (Transgen Biotech, Beijing, China). Ten μL of positive RT-PCR products of the expected sizes were captured with streptavidin Sepharose beads (3 μL; GE Healthcare, Piscataway, NJ ) and binding buffer (47 μL), after which the products were purified using a vacuum prep workstation (Qiagen) with RNase-free water, 70% alcohol, washing buffer, and denaturation buffer in series to form single-stranded DNA. Single-stranded DNA was dispensed into a PSQ 96-well plate (Qiagen) containing 45 μL of annealing buffer containing 10 μM pyrosequencing primer. The plate was heated at 80°C for 2 min and then cooled to room temperature for primer annealing. To obtain sequences, pyrosequencing reactions were performed (PyroMark ID platform, PyroMark Gold reagents, Qiagen) according to the manufacturer’s instructions. When pyrosequencing was performed with the pyrosequencing primers, H5N1 AIVs with serine at the P6 position of the HACS were distinguished by their sequences. Negative RT-PCR products were not submitted for pyrosequencing.

Based on the bioinformatics analysis of the H5 cleavage site, 5 different amino acid residues (G, I, R, S, and a deletion) at the P6 position in all clades of H5 influenza virus were identified. The specificity of the pyrosequencing assay was determined using RNA extracted from rg325G, rg325I, rg325R, rg325S, and rg325*, respectively. These 5 representative H5N1 influenza viruses were generated using reverse genetics as described, each with a different amino acid residue at the P6 position of the HACS to represent different clade preferences. RT-PCR produced clear, single-band products of the expected sizes from the 5 rescued H5N1 influenza viruses, and these positive RT-PCR products were then pyrosequenced. The sequences obtained from the pyrosequencing assays were reverse complementary to the reference strain (i.e., at the P6 cleavage position, the rg325G sequence contained TCC [complementary reverse to GGA], the rg325I sequence contained TAT [complementary reverse to ATA], the rg325R sequence contained TCT [complementary reverse to AGA], and the rg325S sequence contained ACT [complementary reverse to AGT]). The correct sequences were identified at the P6 cleavage position in the 5 rescued H5N1 viruses.

The sequences of RT-PCR–positive samples with expected sizes were also sequenced in both directions at Beijing Tsingke Biological Technology using Sanger sequencing to confirm the results. All pyrosequencing results showed complete agreement with those obtained by Sanger sequencing. Other subtypes of influenza viruses, including H3N2 (A/duck/Anhui/D293/2014), H4N6 (A/mallard/Beijing/10/2016), H9N2 (A/chicken/Hebei/YT/2010), H10N7 (A/mallard/Beijing/27/2011), Newcastle disease virus (LaSota strain), and avian infectious bronchitis virus (H120 strain; species Avian coronavirus) were also assessed in the specificity assessment as controls. No positive PCR products were obtained from these control viruses, indicating the correct analytical specificity of the primers. Our results indicate that the pyrosequencing assay is highly specific.

The rg325S virus was selected as reference virus to evaluate the sensitivity of the pyrosequencing assay. The 50% tissue culture infective dose (TCID₅₀) for rg325S was determined in MDCK (Madin–Darby canine kidney) cells with 10-fold serially diluted virus incubated at 37°C for 48 h. The detection limit of the pyrosequencing assay was evaluated using RNA templates extracted from 10-fold serial dilutions of rg325S virus (1 × 10⁴ to 1 × 10⁰ TCID₅₀/100 µL) for allantoic fluid. The lowest concentration of virus detected by the pyrosequencing assay was 1 × 10¹ TCID₅₀/100 µL. The results of this assay suggest that the pyrosequencing assay is highly sensitive, with a detection limit similar to previous studies.^3,15

To validate the clinical applicability of the pyrosequencing method, 423 clinical specimens, including lungs, brains, livers, tracheas, and oral swabs randomly collected from chickens in Shandong, Beijing, Hebei, and Tianjin provinces of China from 2014 to 2015 were tested. Tissue specimens were homogenized and centrifuged at 8,000 × g for 5 min at 4°C. Supernatants were used for RNA extraction and RT-PCR. No sample tested by this pyrosequencing assay showed the presence of the H5 subtype influenza virus in clinical specimens. Because no positive samples were available from the clinical specimens, 10 positive laboratory infection samples, including lung, spleen, brain, and kidney collected from chickens inoculated with clade 2.3.4.4 H5 AIVs were tested to validate the applicability of this assay. The sequences obtained from the pyrosequencing assay indicated that all 10 positive samples had 1 amino acid deletion at position P6, which is in accord with the sequences obtained from Sanger sequencing.

To compare the RT-PCR H5 detection results with those obtained by virus isolation, the supernatants of the 423 clinical samples were also propagated in the allantoic cavities of 9–11-d-old specific pathogen–free embryonated chicken eggs at 35°C for 72 h for virus isolation. Next, allantoic fluid was harvested and tested for hemagglutination activity. Positive isolates were confirmed as specific subtypes of AIV by hemagglutination inhibition (HI) assay as described previously.¹⁵ Using the conventional virus isolation method and HI assay, 10 laboratory infection samples were identified as clade 2.3.4.4 H5; no sample collected clinically tested positive for the H5 subtype. Fifteen clinical specimens tested positive for hemagglutination activity, and these isolates were confirmed as H9 subtype AIV by the HI assay. Our results show that the pyrosequencing results were in agreement with those obtained using conventional virus isolation and HI assay. However, the traditional methods are time-consuming and insufficient for early monitoring. Our method will provide more rapid detection of the amino acid residue at the P6 cleavage site in HACS when a clinical specimen is identified positively as H5 subtype using traditional methods. Once serine is found in clinical samples, measures should be taken to reduce the risk of the H5N1 AIVs with probable increased pathogenicity in mammals.

Pyrosequencing has been widely used in rapid detection and typing of many pathogens.^3,5 It has also been used in molecular surveillance of influenza viruses by applying it to SNP detection, mutation screening, and reassortant virus identification.⁴ In our study, the pyrosequencing method generated sequence data from H5N1 influenza viruses with different amino acid residues at the P6 position in HACS; RT-PCR and real-time PCR cannot accomplish this task.^1,6,15 Compared with Sanger sequencing, the pyrosequencing assay has similar accuracy, but it is less expensive and time-consuming because the pyrosequencing assay can provide sufficient information to identify the amino acid residue at the P6 position in HACS in a 50-bp sequence instead of the full HA gene by Sanger sequencing. In addition, pyrosequencing takes ~2 h to obtain the sequences of the target genes, but 1–2 d are needed when RT-PCR products were sequenced using Sanger sequencing.

Herein, we developed a pyrosequencing method for detecting H5N1 AIVs possessing serine residue at the P6 cleavage position in the HA gene. Previous studies reported that virulence of influenza virus in mammals is considered to be affected by mutations in multiple viral genes. For instance, PB2 E627K and D701N were known as important determinants for virus replication in mammalian cells.¹⁴ Another study showed that D92E in NS1 increased resistance of H5N1 to tumor necrosis factor–alpha and gamma interferon host responses in vitro and in vivo in swine.¹² Our previous study demonstrated that the amino acid residue at the P6 position of the cleavage site also had a direct correlation with the cleavage efficiency and pathogenicity of H5N1 AIVs in mice. We evaluated the pathogenicity of rg325G, rg325I, rg325R, rg325S, and rg325* and found that HA-325S resulted in the highest pathogenicity in mice.¹⁸ As a result, serine at the P6 position of the HACS may also serve as an indicator of increased pathogenicity of H5N1 AIVs in mammals. Monitoring the emergence and spread of influenza viruses, especially those that pose a serious global health threat because of their high pathogenicity, should be a focus of ongoing influenza surveillance.