Cross-species and mammal-to-mammal transmission of clade 2.3.4.4b highly pathogenic avian influenza A/H5N1 with PB2 adaptations

Abstract

Highly pathogenic H5N1 avian influenza viruses (HPAIV) belonging to lineage 2.3.4.4b emerged in Chile in December 2022, leading to mass mortality events in wild birds, poultry, and marine mammals and one human case. We detected HPAIV in 7,33% (714/9745) of cases between December 2022-April 2023 and sequenced 177 H5N1 virus genomes from poultry, marine mammals, a human, and wild birds spanning >3800 km of Chilean coastline. Chilean viruses were closely related to Peru's H5N1 outbreak, consistent with north-to-south spread down the Pacific coastline. One human virus and nine marine mammal viruses in Chile had the rare PB2 D701N mammalian-adaptation mutation and clustered phylogenetically despite being sampled 5 weeks and hundreds of kilometers apart. These viruses shared additional genetic signatures, including another mammalian PB2 adaptation (Q591K, n = 6), synonymous mutations, and minor variants. Several mutations were detected months later in sealions in the Atlantic coast, indicating that the pinniped outbreaks on the west and east coasts of South America are genetically linked. These data support sustained mammal-to-mammal transmission of HPAIV in marine mammals over thousands of kilometers of Chile's Pacific coastline, which subsequently continued through the Atlantic coastline.

Fig. 1. Unprecedented mortality of marine mammals associated with the HPAIV H5N1 outbreak in Chile.

Documented mortality events of South American sea lions (a) Ñuble Region, (March 19), (b) Arica Region (March 1), (c) Atacama Region, January 10 (d) Tarapaca Region, March 07, (e) Burmeister´s porpoise, Atacama Region (March 29) (f) Chilean dolphin, Maule Region (March, 18) and (g, h) Fin whale, Biobio Region (April 24), in the Pacific coasts of Chile during the H5N1 HPAIV outbreak in Chile during January - April 2023.

Fig. 2. Introduction of H5N1 viruses into Chile.

Time-scaled maximum clade credibility (MCC) tree inferred for the HA segment. The data set includes 177 H5N1 viruses sequenced for this study from marine mammals, poultry, wild birds, and one human case, and 204 background H5N1 viruses collected globally from all species during 2021-2023, downloaded from GISAID on August 25, 2023. Posterior probabilities are provided for key nodes. Branches shaded by location and host species. Posterior density plot, shaded yellow, of the inferred timing (95% HPD) of the introduction of H5N1 into Chile and Peru during the Southern Hemisphere spring (Northern Hemisphere autumn) of 2022. Thick marks in x-axis indicate the separation in the years plotted.

Fig. 3. Interspecies transmission and spread of the Peru/Chile H5N1 virus.

Timescale MCC tree inferred from concatenated genomes. Dataset includes 293 H5N1 viruses from the entire South American clade including 177 viruses sequenced for this study (see Fig. 2). Tree showed the single introduction from Peru (blue) and Chile (light blue and green) and the subsequently spread to Argentina (brown), Brazil (purple) and the South Atlantic islands in the Sub Antarctic Region (light purple). Shaded areas denote specific clusters within the tree.

Fig. 4. Spatial-temporal diffusion of the H5N1 2.3.4.4b lineage in Chile.

A Continuous phylogeographic reconstruction of the spatial spread of the virus inferred from 143 the HA sequences, using a continuous phylogeographic approach implemented using BEAST and visualized using the SERAPHIM package in R. B The timing of the wavefront and dispersal of the H5N1 outbreak in Chile during 2022-2023 visualized using SERAPHIM.

Fig. 5. H5N1 ecology of transmission events in the H5N1 outbreak in South America.

Estimated interspecies transmission events between the 13 affected animal orders in South America. The data set included six orders of aquatic wild birds (yellow), 3 terrestrial wild birds orders (pink), 2 orders of domestic birds (green) and 3 orders of mammals, including the human case, (blue) during the H5N1 outbreak, based on “Markov jump” counts. The width of the arrow is proportional to the estimated number of jumps in that direction as shown by the numbers next to the arrows.

Fig. 6. Global observations of H5N1 D701N and Q591K mutations in the PB2 segment.

ML tree inferred for H5N1 viruses collected from Chile and globally from all host species during January 1, 2021 to September 9, 2024 (n = 12,663 PB2 sequences). Tree is midpoint rooted and branch lengths are drawn to scale. Branches are shaded by avian lineage American and Eurasian. The original introduction of H5N1 from Eurasian into North America in 2021 is labeled. Viruses with D701N (yellow), Q591K (pink) and both D701N/Q591K (light blue) are indicated. Number of marine mammals carrying the mutation are described in front of each cartoon. Inset summarizes the number of viruses carrying the D701N mutation identified globally.

Fig. 7. Amino acid changes and variant analyses in the D701N cluster.

a Amino acid substitutions are listed for the 13 H5N1 viruses in the D701N cluster (Fig. 2), including 11 sequenced for this study and 2 from Peru. These mutations were not observed in wild birds or poultry sequenced in this study, except for the sanderling that is part of the D701N cluster. b Frequency of single nucleotide variants (SNV), and amino acid substitutions for A/H5N1 genomic positions were calculated for the 701N cluster genome segments employing the raw data obtained by Illumina sequencing (when available) or nanopore sequencing. SNV analyses were also performed for all avian segments (Supplementary Fig. 4 and Supplementary Data). Positions are numbered according to the reference genomes OQ352545-OQ352552. Viral isolates are indicated on the left and the are colored according to their geographical origin: Northern Chile (blue), North-Central Chile (brown) or Central Chile (green).