Caractérisation génétique de virus influenza H3N8 équins circulants en France entre 2009 et 2018

International audienceInfluenza viruses are enveloped, segmented negative-strand viruses belonging to the Orthomyxoviridae family. They are the etiologic agents of flu, a respiratory disease affecting many species including equines. Disease severity varies, determined by host factors such as the level of immunity from previous exposure or vaccination, and also viral factors such as virulence and viral load. In horses, influenza virus infects firstly the upper respiratory tract and may be complicated by a secondary infection in the lungs. The flu is caused by influenza type A virus subtypes H3N8 and H7N7, the latter of which has not been detected since the 1980s. Despite the existence of a vaccine, which has an excellent coverage and is very effective, France experienced a few outbreaks of H3N8 since the 2000s, the latest being in 2023. Since 2009, outbreaks were associated with Florida Clade 1 and Clade2 strains facilitated by the horse transportations and competitions, it is highly contagious and spreads around the world with international competitions and exportations, facilitating the global spread of the virus. It is therefore critical to limit these outbreaks, which has an important economic impact for the equine industry. To facilitate the monitoring and management of future outbreaks, obtaining complete genome sequences of the virus are essential to follow the evolution of the virus, as new virulence determinants and escape mutations can appear, potentially rendering the vaccine less effective.We made a genetic characterization of circulating equine influenza H3N8 viruses in France between 2009-2018, to further identify virulence determinants. In particular, we wished to better understand the specificities of four equine viral strains: the A/equine/Paris/1/2018 and A/equine/Beuvron/2009 strains, as well as the two vaccine strains A/equine/Richmond/1/2007 and A/equine/South Africa/4/2003, using Oxford Nanopore Technologies sequencing technology. Sequences of these 4 strains have been compared with the historical strain A/equine/Uruguay/1/1963.First, we analyzed the sequence similarities between the studied strains. We observed that the four sequenced strains show hemagglutinin (HA) and neuraminidase (NA) sequence identity rates of 96-99% and 95-99% between them, respectively, but only 89-91% and 87-90% with Uruguay/1/1963. The sequences of the RNA-polymerase complex subunits PA, PB1 and PB2 as well as that of the nucleoprotein (NP) of these four viruses have at least 97% identity between them. This identity rate drops to 92% with Uruguay/1/1963 strain.We then focused on particular amino-acid substitutions between strains. Several variations have been observed in Paris/1/2018 in the replicative complex that are also found together recently in the American continent when performing a BLAST: 5 in PB2 (G/E65K, A661T, A684T, K699R and I754V) all present since 2016; 4 in PB1 (I114V, V200I, V715A and R754G) all present since 2014, 4 in PA (E59K, E237K, L335I and K367R) all present only since 2018, and 1 in NP (T430I) since 2014.Whilst the NS1 protein of strain Uruguay/1/1963 has a length of 230 amino acids like its human counterparts, the four viruses studied have a truncated NS1 that lacks the 11 C-terminal amino acids for thecontrol of general gene expression. We further observed several variations in the C-terminal of Paris/1/2018 strain that has 3 variations in NS1 (H/Y207N, D/N209I and P212S). These substitutions are also found in USA strains in 2014 to 2021. The consequences of these 3 substitutions on the viral fitness is, however, unknown.Interestingly, we noticed that accessory protein PB1-F2 exhibits the greatest sequence variability between strains. All studied equine strains have at least 3 potential pro-inflammatory markers in their sequence, but can vary between strains. In equine influenza viruses, PB1-F2 has undergone a truncation of 9 amino acids in the C-terminal domain after 1997, producing an 81 amino-acids long polypeptide. This truncation has been reverted in Paris/1/2018. Full-length versions of PB1-F2 (between 87 and 90 amino acids) are known to specifically translocate into mitochondria using their C-terminal region, and exacerbate the inflammatory state. To investigate the contribution of the Paris/1/2018 PB1-F2 to virulence, experiments characterizing the sub-cellular localization of this polypeptide and its truncated homologs will be presented as well as data on their membrane permeabilization capacities.In conclusion, we evidenced that the Paris/1/2018 exhibit specific genomic features that are possibly associated to its virulence and its diffusion despite a high vaccine coverage in France in 2018