Clade 2.3.4.4b H5N1 neuraminidase has a long stalk, which is in contrast to most highly pathogenic H5N1 viruses circulating between 2002 and 2020

Abstract

Since 2020, H5N1 highly pathogenic avian influenza (HPAI) viruses of clade 2.3.4.4b have been rapidly spreading in wild birds but have also caused a large number of mammalian infections and more than 70 known human cases. Importantly, this H5N1 clade has also crossed the species barrier into dairy cattle in the US in late 2023/early 2024. The neuraminidase (NA) protein of the N1 subtype can feature truncations in its stalk domain, which have been identified as putative virulence factors in poultry but seem to have a negative impact on transmission in mammals. Since its emergence, the vast majority of HPAI H5N1 A/goose/Guangdong/1/1996-lineage isolates have featured this truncated version of the NA stalk domain. Here, we report that this changed with the 2020 expansion of clade 2.3.4.4b H5N1 and that the majority of isolates-including the strains circulating in dairy cattle-feature a long NA stalk domain.IMPORTANCEWhile the truncated version of the N1 neuraminidase stalk domain may be associated with increased virulence in poultry, the long version of the stalk domain has been associated with increased transmissibility in mammals. The vast majority of highly pathogenic H5N1 of clade 2.3.4.4b that is currently circulating globally features the long stalk version of the neuraminidase, which may increase the risk for these viruses to become human-to-human transmissible.

Keywords: H5N1; influenza; neuraminidase; neuraminidase stalk.

Fig 1

Overview of long and short N1 NA stalk sequences. (a) Multiple sequence alignment of one representative sequence of H5N1 neuraminidase, downloaded from NCBI Virus (18), from every available year from 1997 to 2024. The insertion in the stalk region that appeared first in 1996 and became the dominant “long stalk” variant from 2021 to 2022 is marked with a red rectangle, predicted N-glycosylation sites included in this region are highlighted in yellow. As a reference, the A/goose/Guangdong/1/1996 sequence was used. NCBI accession numbers of the used sequences are as follows, starting from 1997: ACZ45121.1, ACZ45106.1, ACZ45112.1, ACZ45114.1, ACA47584.1, BAM85821.1, AGH30708.1, ABU99043.1, ABV24001.1, AGK72265.1, AGJ73095.1, AUD40562.1, AER09772.1, BAM62219.1, AIZ06790.1, AGM16067.1, APG38584.1, AMX74824.1, ARX75171.1, QHQ71777.1, QEP94697.1, QIC52185.1, UMW88849.1, UTS97654.1, WHA13083.1, WWE95248.1, and XDL05563.1. Alignment was done with Clustal Ω (19). (b) Percentage of NA sequences from H5N1 viruses per year that contain the long stalk N1 (as defined in the "Methods" section in the Supplemental Material). (c) Percentage of NA sequences from H5N1 viruses from clade 2.3.4.4b that contain the above long stalk N1, per host. Numbers at the base of the bars indicate the total number of sequences, and numbers on top indicate the percentage having the long stalk. For avian species, the search was limited to samples collected from 1 January 2024, and for cattle sequences, from 1 January 2023 (downloaded on 15 July 2024 or 19 July 2024). Poultry was defined as any sequences containing the words “chicken,” “turkey,” “domestic duck,” “domestic goose,” “domestic,” or “poultry” in their header. All other avian samples were classified as wild birds.

Fig 2

A phylogenic tree created using H5N1 NA sequences with long (blue squares) or short stalk (pink circles). The tree was rerooted to an N4 sequence (A/mallard/Sweden/24/2002, EPI267067 (GISAID)) included in the analysis (not shown on the tree). The orange inset zooms into the only branch (marked with orange) where both long- and short-stalk sequences appear, along with the ancestral A/goose/Guangdong/1/1996 sequence (purple circle). The “1997 stalk” represents a different branch with a distinct deletion compared to what is observed from 2000. The tree was created using the FastTree one click workflow of the Ngphylogeny.fr (21) server and visualized with iTOL (22). The groups marked with years indicate that the branch contains sequences from that time period, with one or two outliers. The tree scale and branch lengths refer to the number of substitutions per site.