A Dual Motif in the Hemagglutinin of H5N1 Goose/Guangdong-Like Highly Pathogenic Avian Influenza Virus Strains Is Conserved from Their Early Evolution and Increases both Membrane Fusion pH and Virulence

Abstract

Zoonotic highly pathogenic avian influenza viruses (HPAIV) have raised serious public health concerns of a novel pandemic. These strains emerge from low-pathogenic precursors by the acquisition of a polybasic hemagglutinin (HA) cleavage site, the prime virulence determinant. However, required coadaptations of the HA early in HPAIV evolution remained uncertain. To address this question, we generated several HA1/HA2 chimeras and point mutants of an H5N1 clade 2.2.2 HPAIV and an H5N1 low-pathogenic strain. Initial surveys of 3,385 HPAIV H5 HA sequences revealed frequencies of 0.5% for the single amino acids 123R and 124I but a frequency of 97.5% for the dual combination. This highly conserved dual motif is still retained in contemporary H5 HPAIV, including the novel H5NX reassortants carrying neuraminidases of different subtypes, like the H5N8 and the zoonotic H5N6 strains. Remarkably, the earliest Asian H5N1 HPAIV, the Goose/Guangdong strains from 1996/1997, carried 123R only, whereas 124I appeared later in 1997. Experimental reversion in the HPAIV HA to the two residues 123S and124T, characteristic of low-pathogenic strains, prevented virus rescue, while the single substitutions attenuated the virus in both chicken and mice considerably, accompanied by a decreased HA fusion pH. This increased pH sensitivity of H5 HPAIV enables HA-mediated membrane fusion at a higher endosomal pH. Therefore, this HA adaptation may permit infection of cells with less-acidic endosomes, e.g., within the respiratory tract, resulting in an extended organ tropism. Taken together, HA coadaptation to increased acid sensitivity promoted the early evolution of H5 Goose/Guangdong-like HPAIV strains and is still required for their zoonotic potential.IMPORTANCE Zoonotic highly pathogenic avian influenza viruses (HPAIV) have raised serious public health concerns of a novel pandemic. Their prime virulence determinant is the polybasic hemagglutinin (HA) cleavage site. However, required coadaptations in the HA (and other genes) remained uncertain. Here, we identified the dual motif 123R/124I in the HA head that increases the activation pH of HA-mediated membrane fusion, essential for virus genome release into the cytoplasm. This motif is extremely predominant in H5 HPAIV and emerged already in the earliest 1997 H5N1 HPAIV. Reversion to 123S or 124T, characteristic of low-pathogenic strains, attenuated the virus in chicken and mice, accompanied by a decreased HA activation pH. This increased pH sensitivity of H5 HPAIV extends the viral tropism to cells with less-acidic endosomes, e.g., within the respiratory tract. Therefore, early HA adaptation to increased acid sensitivity promoted the emergence of H5 Goose/Guangdong-like HPAIV strains and is required for their zoonotic potential.

Keywords: H5N1; HA; HPAIV; hemagglutinin; influenza virus; virus evolution.

FIG 1

Alignment of consensus sequences of the HA proteins from LPAIV, TG05, HPAIV, and R65. For analysis, we surveyed the fluDB (38) and GISAID (https://www.gisaid.org/) public databases and included 4,188 full-length aligned sequences to generate the consensus HA sequences from LPAIV (803 entries) and HPAIV (3,385 entries).

FIG 2

Frequencies of the dual motif 123R/124I (A) and the N-glycosylation motif spanning residues 156 to 158 (B) in H5 HPAIV and generated recombinant viruses (C). Shown are frequencies of amino acids at positions 123 and 124 (A) or an N-glycosylation motif at positions 156 to 158 (B) within H5 HA of HPAIV and LPAIV among 4,188 protein sequences downloaded from the fluDB (38) and GISAID (https://www.gisaid.org/) public databases on 24 July 2017. Positions refer to the R65 HA, including the signal peptide.

FIG 3

Growth curves. We inoculated DF1 cells with virus at a multiplicity of infection of 10⁻³ in duplicate and determined the average titers at the respective time points by plaque assays. In panels B and C, the growth curves for R65, R65-HA1_TG05poly/HA2_R65, and R65-HA1_R65/HA2_TG05 are identical to those from panel A and are included for comparison (dotted lines).

FIG 4

HA1 R123S and I124T exchanges abolish or reduce virulence in chicken. We infected the birds oculonasally with 10⁵ PFU virus and added contact animals on day 1 p.i. Daily individual clinical scores were 0 for healthy, 1 for ill, 2 for severely ill, and 3 for dead. CS, total clinical score for the whole group of the primarily infected (inoculated) chickens.

FIG 5

HA1 exchange R123S abolishes virulence and reduces virus load in organs of mice. (A) Average body weights of 4-week-old female BALB/c mice inoculated intranasally with mock (PBS), R65, or R65-HA_R123S (n = 5) at a dosage of 10⁴ PFU. Animals with weight losses of 25% or more were euthanized. All R65-infected mice died, whereas all R65-HA_R123S-infected animals remained unaffected throughout the experiment. (B) Organ titers, determined by plaque assays, from entire lungs, hearts, and brains taken on day 3 after intranasal inoculation with 10⁴ PFU of R65 or R65-HA_R123S. If necessary, titration began with undiluted homogenates.

FIG 6

Fusion activation pH and stability. (A) Decreased fusion activation pH and virus inactivation pH of HA chimeras and point mutants. For each virus, the left end of the bar indicates the fusion activation pH (the highest pH at which syncytia formed), whereas the right end represents the virus inactivation pH (the highest pH at which no cell infection was detectable). (B) Thermal stability of HA chimeras and point mutants. After incubation of the virus at the indicated temperatures for 30 min, we determined the HA titers at room temperature.

FIG 7

Virulence in chickens increases with fusion activation pH. For each virus, the average clinical score in chickens is shown, along with the fusion activation pH determined by syncytium formation assays. Clinical scores are 0 for healthy, 1 for ill, 2 for severely ill, and 3 for dead. *, clinical scores of these viruses were determined previously (18).

FIG 8

Mechanistic implications for membrane fusion. (A) Prefusion conformation (PDB accession number 1JSM) (49) (left) and postfusion conformation (PDB accession number 4NKJ) (87) (right) of a hemagglutinin trimer. The HA1 subunit is shown in a transparent surface representation, and the HA2 subunit is shown in a cartoon representation. The different structural elements of a single HA2 protomer are highlighted in different colors. The rectangular outline indicates the region of close-ups in panels B to F. (B) The B loop conformation in a prefusion-stabilized mutant of A/Japan/305/1957 H2 at pH 8.1 (PDB accession number 3QQB) (47). (C) The same mutant at pH 5.3 (PDB accession number 3QQO) (47). (D) LPAIV A/duck/Singapore/3/1997 H5 at pH 7.5 (PDB accession number 1JSM) (49). (E) HPAIV A/Chicken/Hong Kong/YU562/01 H5 at pH 6.6 (PDB accession number 3S13) (37). (F) The same protein in another crystal form grown under identical conditions (PDB accession number 3S12) (37).