Migratory birds in southern Brazil are a source of multiple avian influenza virus subtypes

Background There is insufficient knowledge about the relation of avian influenza virus (AIV) to migratory birds in South America. Accordingly, we studied samples obtained over a 4‐year period (2009‐2012) from wild birds at a major wintering site in southern Brazil. Methods We obtained 1212 oropharyngeal/cloacal samples from wild birds at Lagoa do Peixe National Park and screened them for influenza A virus by RT‐PCR amplification of the matrix gene. Virus isolates were subjected to genomic sequencing and antigenic characterization. Results Forty‐eight samples of 1212 (3.96%) contained detectable influenza virus RNA. Partial viral sequences were obtained from 12 of these samples, showing the presence of H2N2 (1), H6Nx (1), H6N1 (8), H9N2 (1), and H12N5 (1) viruses. As H6 viruses predominated, we generated complete genomes from all 9 H6 viruses. Phylogenetic analyses showed that they were most similar to viruses of South American lineage. The H6N1 viruses caused no disease signs in infected ferrets and, despite genetic differences, were antigenically similar to North American isolates. Conclusions Lagoa do Peixe National Park is a source of multiple AIV subtypes, with the levels of influenza virus in birds being highest at the end of their wintering period in this region. H6N1 viruses were the predominant subtype identified. These viruses were more similar to viruses of South American lineage than to those of North American lineage.

different flyways. 4 Every year, millions of wild birds from different North American flyways converge at wintering sites in Brazil. The Lagoa do Peixe National Park is recognized as the most important of these stopovers and acts as a wintering site for North American shorebirds. [5][6][7] The birds usually remain in Brazil from September to May, and they depend on this habitat to replenish the energy expended in their winter migration. 8 Lagoa do Peixe National Park is also important to resident birds and to South American migrant birds, 9 suggesting that there is great potential there for the spread and genetic reassortment of influenza A viruses. Despite the commingling of birds from major North and South American flyways, there is a marked phylogenic separation of some influenza viruses residing within them. Based on genetic sequence, there are 2 major groups of influenza viruses in the Americas, one made up of viruses from North and South America, and the second only with viruses from South America. 10 Influenza A viruses are composed of 8 RNA segments and are subtyped according to their hemagglutinin (HA) and neuraminidase (NA) antigenic glycoproteins. Evidence suggests that viruses representing the 16 HA and 9 NA subtypes exist in harmony with their natural reservoirs, causing little or no disease. 11 AIVs in Eurasia and Australia appear to have diverged from those in North America, presumably as a result of different flyways being established and a consequent lack of contact between birds and therefore viruses, from the different continents. 12 Similarly, the HA and NA genes of AIVs in Argentina differ from those of viruses circulating in North America and Eurasia. 13 The main subtypes of AIV consistently reported to circulate in other countries [14][15][16][17][18][19] have not been reported in Brazil. However, an H11N9 influenza virus was recently identified for the first time in South American migratory birds in the Amazon region, and this virus showed great similarity to viruses of North American, rather than South American, lineages. 20 Compared to surveillance activities in other regions of the world, influenza virus surveillance in South American wildlife has been limited. It remains unclear whether the North American and South American AIV lineages converge and whether migrating wild birds transport AIVs of South American lineage to the United States. 21 Much remains unknown about AIV ecology in Brazil, including the identity of the main virus carrier species and the extent to which wild birds support virus spread between North and South America. Here, we studied samples obtained from wild birds in southern Brazil over a 4-year period to determine the extent of AIV circulation, the major host species of AIV, and the nature of any viruses present.

| Ethics statement
The study was conducted in strict accordance with the guidelines of the

| Virus isolation and subtyping
Virus was isolated by standard methods in 10-day-old embryonated chicken eggs, and viral RNA was amplified from samples positive for AIV by real-time RT-PCR. 23,24 The Flu DETECT Avian Influenza Virus Type A Test Kit (Synbiotics, San Diego, CA, USA) was used to quickly confirm the presence of AIV protein within each sample. 25 Hemagglutination and neuraminidase inhibition tests were performed with a panel of polyclonal reference sera to identify the influenza subtypes as recorded in Table 1. [26][27][28]

| Phylogenetic analysis
The complete sequences of all segments of each isolate were compared using the Basic Local Alignment Search Tool (BLAST) to identify the most closely related sequences available in public databases. 35 The nucleotide sequences of each gene segment were aligned using clc genomics workbench version 7.6 (Qiagen, Aarhus, Denmark), and the percentage nucleotide similarity was calculated using the MegAlign program of the

| Ferret studies
To prepare post-infection ferret antiserum, ferrets aged approximately 8 months (Triple F Farms, Sayre, PA, USA), previously T A B L E 1 Forty-eight samples testing positive for influenza A virus. The 17 samples in which the presence of a specific subtype was confirmed are in bold. Nine H6N1-subtype viruses were confirmed, along with one virus of each of the H6Nx, H2N2, H12N5, and H9N2 subtypes

| Virus detection
We examined 1212 samples from wild birds. Cloacal and oropharyngeal swabs from each bird were collected and combined to facilitate the analysis (Table 2). Each year of the study featured 2 distinct events:

| Genetic and antigenic analyses of isolated H6N1 viruses
In our genetic analysis, we were able to generate complete genomes from 9 samples. Eight of these genomes were obtained from swab material, and only one was obtained from an egg isolate.

| DISCUSSION
In our 4 years of sampling at this important migratory bird wintering site, we found evidence of AIV circulation in consecutive years, with Bootstraps values >50% were obtained in the analysis of 1000 replicates and are presented at the branching points. The phylogenetic trees were constructed using maximum likelihood methodology. Detailed phylogenetic trees with virus designations can be found in Figure S1 in the Supporting Information T A B L E 3 HI analysis of North American and South American H6N1 viruses. The HI assays were conducted in triplicate, and the titers are reported as the geometric mean. An HI titer of 40 or higher is indicative of seroconversion Kiscporski, Roberta C. Piuco, and Suzana Seibert.

CONFLICT OF INTEREST
None of the authors have a conflict of interest.

BIOGRAPHICAL SKETCH
Dr. Jansen Araujo has a Ph.D. in Microbiology and is the coordinator of the field team in the Virology Laboratory at the Biomedical Science Institute at the University of São Paulo. His research interests include eco-epidemiological studies of wildlife and emerging infectious diseases with zoonotic potential in Brazil.

STATEMENT
Wild birds are the natural reservoir for avian influenza viruses (AIVs).
Brazil provides stopover and wintering sites for migratory birds. We wished to investigate the potential for influenza A viruses to be transported by birds that migrate annually along with resident species and to determine the subtypes involved and their potential for reassortment. In Brazil, data about AIV subtypes are scarce, and few groups have worked with AIV in wild birds. At present, only AIV H11N9 has been isolated from the Amazon region (in 2014). This study was conducted to elucidate which AIV subtypes were present at an important stopover and wintering site of wild birds in Brazil and the genetic relatedness between viruses in the Northern and Southern Hemispheres.
In addition to performing isolation and next-generation sequencing (NGS), we used a ferret model to generate influenza A antiserum to compare antigenic differences in influenza viruses from the Northern and Southern Hemispheres.