First detection of an H5N2 subtype of Influenza A virus detected in Charadrius collaris from the Brazilian Pantanal

Abstract

The migratory bird orders Anseriformes and Charadriiformes are natural reservoirs of avian influenza virus (AIV) and key contributors to its global dissemination. Stopover and wintering areas used by these species are critical hotspots for virus transmission, underscoring the need for routine surveillance. The Pantanal biome in central South America—an extensive floodplain crossed by two major intercontinental migratory routes—hosts numerous species of migratory birds. Given the ecological relevance of this biome and the health risks associated with AIV, this study aimed to investigate AIV circulation in the region. A total of 1,108 orotracheal and cloacal swab samples from 157 bird species were analyzed using RT-qPCR targeting the matrix gene of AIV. Influenza A virus RNA was detected in Charadrius collaris. Subsequent sequencing of the HA and NA genes revealed 94% similarity with H5N2 strains previously identified in Colombian Anatidae in 2011. Additional gene segments (NP;NS;PA;M;PB1;PB2) showed similarity to strains from America and Asia continents. This is the first report of an AIV-infected bird in the Brazilian Pantanal and confirms the presence of a low-pathogenic H5N2 subtype. The findings highlight the role of migratory birds in viral dissemination inland and the potential role of resident species as local reservoirs.

Supplementary Information

The online version contains supplementary material available at 10.1038/s41598-026-42819-y.

Introduction

Avian influenza virus (AIV) is among the most important causes of infections in domestic and wild birds^1,2. In addition to its elevated transmissibility and worldwide distribution, AIV has an important zoonotic potential. For this reason, AIV notification to the World Organization for Animal Health (WOAH) is mandatory as its presence poses major challenges for commercial poultry farming, frequently associated with high morbidity and mortality rates^3,4. Different AIV subtypes are classified according to the antigenic characteristics of the envelope proteins Hemagglutinin (H) and Neuraminidase (N)^5,6.

Epidemics due to highly pathogenic influenza viruses (HPAIV) cause socioeconomic problems, represent major threats to animal and human health worldwide since these viruses are capable of breaking interspecies barriers (spillover)^7,8. The H5 and H7 AIV subtypes are mainly responsible for zoonotic infections. According to the WOAH Terrestrial Manual⁹, H5 and H7 subtypes are considered notifiable due to their potential to become highly pathogenic. In addition to causing infections in Anseriformes and Charadriiformes (natural hosts) and other bird species, they can infect mammalian hosts for example^7,10,11. In particular, low pathogenic AIVs subtypes (LPAIV) H5 and H7 are also susceptible to mutation, mainly after rapid propagation in commercial flocks, becoming highly pathogenic and virulent variants^5,12. These subtypes have also been associated with cross-species transmission events at the wild-domestic interface¹⁰.

In their natural hosts, AIV exhibit tropism for the epithelial tissues of the respiratory and gastrointestinal systems, where the highest rates of replication and excretion of viral particles occur^7,13. Therefore, the fecal–oral route comprises the main form of propagation in the wild-domestic transmission cycle and between species from different geographic regions^14,15. In this sense, migratory waterbirds play a crucial role in viral dissemination over long territorial distances, and are related to the spread of several types of AIV^16–18. Waterbirds contribute significantly to viral amplification in the environment, excreting high amounts of viral particles in their feces for more than 20 days, contaminating water, food, and soil, especially in humid environments during their flight routes^7,8,19,20.

The current HPAIV (H5N1) panzootic outbreak began in 2020 arriving in Brazil in May 2023 through the migratory Atlantic route, causing mortality in wild and domestic birds throughout the Southern Hemisphere^21,22. Furthermore, the risks of additional epidemics by other avian subtypes exist. For example, the low pathogenicity H5 subtype of AIV has continued to be active in domestic birds in North^22,23 and Central America²¹ for more than 2 decades, with potential risk for mutations into highly pathogenic subtypes as occurred in México and Taiwan^24,25.

Given this scenario, the present study investigated the presence of AIV in wild birds in the Northern Pantanal region at Mato Grosso State, Brazil, focusing important sites of stopover and wintering of migratory birds coming from different areas of North, Central, and South America continents. With peculiar physical characteristics, the Pantanal is the largest floodplain in the world, covering approximately 150,000 km², placed in the center of South America²⁶. Due to its geographical position and the presence of beaches, rivers and large lakes used by migratory birds coming from the Central Brazil and Central Amazon/Pantanal routes^27–29, the Pantanal is considered a hotspot area for the occurrence of spillover²⁶. In addition to the massive presence of migratory birds, this region is also known for harbor regional birds, which even migrate to other areas of Brazil. In these locations, there is a vast diversity of wildlife close to peri-urban areas, where close contact between wildlife, domestic fauna, and humans is often observed.

Results

Migratory bird sites in the Pantanal and characteristics of the studied birds

The capture campaigns resulted in 1,108 birds sampled, representing 157 species across 17 orders. Sampling was conducted during 17 campaigns between 2021 and 2023 in four municipalities of Mato Grosso State: Cuiabá, Nobres, Poconé, and Santo Antônio do Leverger. These areas are either located within or in close proximity to the Pantanal biome (Cuiabá and Nobres). Migratory bird sites in the bays of Chacororé and Sinhá Mariana, Mimoso field and on the banks of the rivers Cuiabá, Coxipó and São Lourenço were defined as collection points (Table 1, Fig. 1).

Table 1.

Geographic coordinates of wild birds’ collection sites, their municipalities and localities in the Pantanal biome, Mato Grosso State, Brazil.

Municipality	Location	Geographic coordinates
Cuiabá	1. Coxipó River	− 15.615256; − 56.010385
Nobres	2. Cuiabá River	− 14.630146; − 56.157662
Poconé	3. Cuiabá River	− 16.944486; − 56.637640
Santo Antônio do Leverger	4. Bay of Chacororé	− 16.263843; − 55.859076
	5. Bay of Sinhá Mariana	− 16.340140; − 55.899194
	6. Mimoso field	− 16.211385; − 55.807308
	7. Mimoso field	− 16.338971; − 55.899045
	8. São Lourenço River	− 16.754653 − 55.567378

Fig. 1.

Wild bird collection point in the northern region of the Brazilian Pantanal. Point #4, identified as Chacororé Bay, is the location where the bird Charadrius collaris captured on one of the bay’s beaches was sampled. All map elements and illustrations are original and created by the author.

From each bird, one oropharyngeal and one cloacal swab were collected and stored in separate cryotubes. A third cryotube, referred to as the “mixed cryotube”, was used to store an additional oropharyngeal and cloacal swab from the same individual. The individual (non-mixed) swabs were tested separately whenever the mixed sample tested positive, in order to confirm the result and determine the presence of the virus in each mucosal site.

Based on species-level taxonomic identification, the occurrence status of each species in Brazil was classified according to Pacheco et al.³⁰. A total of 1036 (93.5%; 1,036/1,108) resident/migratory birds with reproductive evidence in the country were identified. Birds in this category undertake migratory movements within the national territory.

A total of 72 (6.5%; 72/1,108) intercontinental migratory birds were captured, of which 64 belonged to the order Charadriiformes, 7 to the order Passeriformes, and 1 to the order Accipitriformes. Sixty-five (5.86%; 65/1,108) individuals were identified as non-breeding visitors from the North (VI (N)), the Nearctic region, the Caribbean, or the extreme north of South America; six (0.5%) as non-breeding visitors from areas west of Brazilian territory (VI (W)); and only one (0.09%) non-breeding visitor from areas south of Brazil (VI (S)), such as Uruguay or Chile. Supplementary Tables 1 and 2 presents the intercontinental migratory birds captured in the present study as well as the occurrence records in Brazil.

Fecal samples collection

A total of 94 fecal samples were collected from the environment, 73 of then belonged to a group of birds from the order Charadriiformes; family Laridae, comprising 35 samples from the species Rynchops niger and 28 from Phaetusa simplex; family Charadriidae, with nine fecal samples from a group of Charadrius collaris; family Recurvirostridae, one sample from a group of Himantopus melanurus. Fecal samples from 17 groups of the Anseriformes order were also collected, belonging from family Anatidae and species Dendrocygna autumnalis. Three fecal samples came from a group of Syrigma sibiliatrix that belongs to the family Ardeidae and order Pelecaniformes. One fecal sample was from a group of Nannopterum brasilianum, which belongs from the Suliformes order and Phalacrocoracidae family.

TaqMan RT-qPCR assay results

Of the 1108 mixed swab samples evaluated, one from the bird MT1403 presented Ct value of 28 for detection of the AIV Matrix gene. The positive bird belongs to the order Charadriiformes, species Charadrius collaris, popularly known as the collared plover. Additional RT-qPCR analyses using individual samples of orotracheal, and cloacal swabs resulted in amplifications with Cts values of 24.4 and 38.2, respectively. This specimen was captured in August 2023 in the Bay of Chacororé by active capture (Fig. 1). There was no detection of AIV in stool samples collected from the environment.

Phylogenetic analysis

Next Generation Sequence of sample MT1403 generated a total of 2,557,072 reads. Full-length sequences of all 8 segments were recovered with robust coverage metrics varying between 317.78 and 7581.34 (Supplementary Table 3). The nucleotides sequences generated from the HA and NA genes demonstrated that both presented ~ 96% similarity with sequences generated from wild anatids (A/white-faced whistling duck/Colombia/1/2011(H5N2) in Colombia 2011 (Genbank codes: CY207181 and CY207183 respectively). Values of 93% were related to subtypes detected in North America, and there was not relationship with HPAIV H5N1 subtype in the Americas. Analysis of the deduced HA sequence showed a cleavage site consistent with a low pathogenic AIV (VPQKETRGLF). Our cleavage site sequence was 100% similar to several low pathogenic H5N2 isolates among other H5 subtypes detected in domestic and wild birds in North America, including a sample detected in chicken from Mexico in 2006 (GenBank AIN25329). Comparing with the last H5N2 isolate detected in Colombia, our sequence showed divergence in amino acid residue 339, deduced as Lysine (K), while the Colombian isolates presented the residue Arginine (R) (GenBank AGI02467, AGI02478 and ANJ78109). Table 2 presents similarity values of the amino acid sequences deduced from our H5N2 isolate (A/collared plover/Pantanal1403/Brazil/2023 (H5N2)) and different samples of H5 and N2 subtypes described in the Americas in the last 20 years.

Table 2.

Similarity between the amino acid sequence deduced from the Hemagglutinin (HA) and Neuraminidase (NA) proteins of the H5N2 subtype detected in Charadrius collaris from the Brazilian Pantanal compared to isolates detected in different regions of America continent.

Gene	GenBank (Description)	Similarity
HA	AGI02467 (A/white-faced whistling duck/Colombia/1/2011(H5N2)	96.2%
	AGI02478 (A/black-bellied whistling-duck/Colombia/1/2011(H5N2))	96.2%
	ANJ78109 (A/black-bellied whistling-duck/Colombia/1/2011(H5N2))	96.0%
	ACV90743 (A/mallard/Louisiana/476670-4/2007(H5N2))	95.2%
	AGC73810 (A/mallard/Ohio/11OS2006/2011(H5N2))	95.2%
NA	ANJ78100 (A/white-faced whistling duck/Colombia/1/2011(H5N2))	95.3%
	AGE03002 (A/American green-winged teal/Wisconsin/10OS2955/2010(H5N2))	96.1%
	AGK42673 (A/mallard/Minnesota/Sg-01040/2008(H3N2))	96.8%
	ACY79662 (A/wild bird/Wisconsin/439436/2006(H5N2))	96.5%
	AAP57557 (A/turkey/CA/D0208652-C/02(H5N2))	96.5%

The HA and PB2 segments were also analyzed for the presence of known mammalian-adaptive molecular markers. In the HA gene, the amino acid residues at positions 226 and 228 were glutamine (Q) and glycine (G), respectively, which are characteristic of avian-type receptor binding. Similarly, the PB2 gene exhibited glutamate (E) at position 627 and aspartate (D) at position 701, consistent with avian adaptation.

Phylogenetic analysis of nonstructural genes of the H5N2 virus detected in our study revealed different patterns of genetic proximity. Most of the segments showed greater similarity with North American strains isolated from wild anatids (Genbank: CY206792 and MT824854, respectively). In contrast, the NP gene showed greater relationship with H5N2 variants isolated from Mexican, Korean and Alaskan birds (Genbank: KC669390, CY098231 and CY206792, respectively). As well as PA segment demonstrated genetic proximity with groups of influenza viruses of Asian origin (Genbank: ON513994). Table 3 shows the different similarity values between the nucleotide sequences of the eight genomic segments of the H5N2 isolate of the present study compared to different isolates sequenced around the world.

Table 3.

Similarity value of the eight genomic segments of the H5N2 subtype of AIV detected in Charadrius collaris captured in the Brazilian Pantanal compared to isolates detected in different regions of the world.

Gene (segment)	GenBank (Description)	Similarity
PB2 (1)	CY004345 (A/pintail/Alberta/49/2003(H12N5)	94.0%
PB1 (2)	MT824854 (A/American Green-winged Teal/Missouri/17OS3221/2017(H3N8)	98.0%
PA (3)	ON513994 (A/Bean Goose/South Korea/KNU2021-25/2021(H5N8))	93.3%
HA (4)	CY207181 (A/white-faced whistling duck/Colombia/1/2011(H5N2)	93.7%
NP (5)	CY206792 (A/mallard/Southcentral Alaska/12ML01055/2014(H1N1)	96.7%
NA (6)	CY207183 (A/white-faced whistling duck/Colombia/1/2011(H5N2)	94.0%
M1 (7)	GU051360 (A/mallard/Minnesota/319/1998(H6Nx)	95.2%
NS (8)	CY033336 (A/northern shoveler/California/HKWF1046/2007(H3N5)	97.0%

Phylogenetic analysis of the Influenza A (H5N2) virus obtained from Pantanal showed that its HA clustered (Bootstrap > 70) with viruses isolated from birds sampled in Colombia in 2011 and Canada in 2010; and distant from viruses isolated in Asia, the United States and Mexico (Fig. 2). Our isolate also shared over 92% HA identity with H5N3 AIV strains previously reported in Argentina (GenBank: PP502509) and Chile (GenBank: KX185908). The NA segment was positioned in an isolated branch (Bootstrap > 70), but close to a clade consisting of H5N2 samples from Colombia and several other isolates detected in the United States and Central America (Fig. 3).

Fig. 2.

Maximum-likelihood phylogenetic tree of the hemagglutinin (HA) segment based on an alignment of 106 sequences with 1716 nucleotides in length. Bootstrap was set to 1000 for statistical support. The tree was inferred using the GTR + F + I + G4 nucleotide substitution model, selected by ModelFinder. Tip labels are color-coded by geographic origin: blue for North America, green for Eurasia, violet for South America, and red for Brazil—Pantanal. Branch support values above 70 are indicated by light blue circles. The scale bar represents the number of substitutions per site.

Fig. 3.

Maximum-likelihood phylogenetic tree of the neuraminidase (NA) segment based on an alignment of 106 sequences with 1410 nucleotides in length. Bootstrap was set to 1000 for statistical support. The tree was inferred using the GTR + F + G4 nucleotide substitution model, selected by ModelFinder. Tip labels are color-coded by geographic origin: blue for North America, green for Eurasia, violet for South America, and red for Brazil—Pantanal. Branch support values above 70 are indicated by light blue circles. The scale bar represents the number of substitutions per site. The other non-structural segments presented different positions. Segments 1(PB2), 2 (PB1) and 5 (NP) grouped into clades consisting of different subtypes occurring in the American continent. Segment 3 (PA) remained in a single branch but close to different subtypes samples detected in Asia and the western coast of the United States and Canada, as well as segment 7 (MA) that was positioned in a distinct branch but in a clade consisting of subtypes detected in North America. Finally, segment 8 (NS) grouped into a clade consisting of the H5N2 subtypes detected in Colombia in 2011. The phylogenetic trees for the non-structural segments of AIV are presented in Supplementary Figs. 1 to 6.

Discussion

This study conducted the most extensive survey of AIV in wild birds in the Brazilian Pantanal. There were more than 1000 swab samples from wild birds collected during a period of two years in 17 field work. Although our sampling design does not allow for prevalence estimation, it reflects the inherent challenges of conducting structured surveillance in wild avian populations within remote and dynamic ecosystems such as the Pantanal, where bird densities and movement patterns are highly variable and largely undocumented. Nevertheless, the scale, temporal coverage, and ecological targeting of this study significantly contribute to the understanding of AIV circulation in inland South America.

Seventy-three birds with intercontinental migratory habits were captured, of which 64 belonged to the order Charadriiformes, seven to Passeriformes, and two to the order Accipitriformes. According to data from Brazilian ornithological records, 68 of the 73 species of migratory birds have been recorded in all Brazilian biomes and can be found in practically all states of the country (Supplementary Tables 1 and 2), this is why the Pantanal biome is considered a hotspot environment as it brings together migratory birds from different locations, in addition to being part of two intercontinental routes for migrant birds³¹. The geographical position of the state of Mato Grosso in the geodetic center of South America, combined with the Pantanal floodplain, provides important stopover points for migratory birds, allowing the exchange of pathogens between migratory and resident avifauna³².

The capture of migratory birds (73 individuals), predominantly from the Nearctic region, the Caribbean and the extreme north of South America, reinforces the importance of the Pantanal for long-distance migratory birds. As shown in Supplementary Table 2, except for two passerine species (Inezia inornata and Serpophaga subcristata), which were captured at the beginning of the dry season in the Pantanal (May), all Anseriformes and Charadriiformes species were captured during the second half of the year. As previously reported by ornithological studies in the Pantanal, Nearctic migratory birds arrive in the region during their southward migration in early austral spring, coinciding with the local dry season. This period favors the exposure of riverbanks, sandbars, and shallow lagoons—ideal environments for feeding and resting during migration^27,33.

Since most migratory birds were captured after their arrival from the Nearctic region, the detection of at least one AIV subtype was expected. On the other hand, the return of migratory birds to the Northern Hemisphere occurs preferentially along the Brazilian Atlantic coast, as this period coincides with the flooding season in the Pantanal, when the wintering fields are submerged. Although sampling did not follow a systematic design and resulted in fewer samples than expected the timing of captures during peak migratory periods increased the likelihood of detecting influenza-infected birds.

This survey identified for the first time a wild bird infected by an AIV in the Brazilian Pantanal. Through phylogenetic analyses, it was possible to define the virus detected as subtype H5N2. A previous survey conducted on wild birds by Pinto et al.³² in the same region in 2014 did not detect birds infected with AIV. The experiment was carried out in a similar capture method to that adopted in this experiment, however, analyzing a smaller number of samples becoming from a reduced sampling effort and mainly testing Passeriformes birds. Therefore, our result reinforces that the monitoring of AIV in wild birds in the Pantanal region needs to be conducted in several sampling sites with a minimum effort of 12 to 20 h per campaign, focusing mainly on the main reservoirs of the virus, such as birds of the orders Anseriformes and Charadriiformes⁸. Additionally, any mortality outbreaks in other avian orders should be part of the AIV monitoring service in the Brazilian Pantanal and therefore need to be tested.

Because there are no reports of avian mortality or confirmed AIV infection in wild birds in the northern Pantanal, we hypothesized that, with sufficient sampling effort, active viral circulation may be occurring without overt clinical signs, thereby increasing the likelihood of AIV detection through active surveillance. A serological survey conducted by Bourscheid et al.³⁴ in backyard poultry raised near migratory bird sites in the Pantanal revealed the occurrence of AIV seropositive poultries, however without antibodies titers detection of H5 subtype among these individuals. Therefore, the detection of the H5N2 subtype in wild birds in the Pantanal is pertinent because it demonstrates that this viral subtype is active even after 10 years without reports in wild birds in the Americas, since its last description was in 2011 infecting anatids in Colombia³⁵.

In addition to the findings previously described, this study reports the first detection of H5N2 in wild birds in Brazil. The amino acid sequences of the HA and NA proteins showed 96% identity to strains detected in anatids from Colombia in 2011. Our isolate also shared 92.5 and 92.2% HA identity with H5N3 strains from Argentina and Chile, respectively. However, phylogenetic analysis revealed that the Brazilian H5N2 isolate clustered within a well-supported monophyletic clade alongside older Colombian H5N2 viruses and was clearly divergent from the Argentinean and Chilean H5N3 strains (Supplementary Fig. 7). Although the moderate identity observed in the HA segment may reflect ancestral relationships or conserved regions, the distinct phylogenetic placement suggests independent evolutionary histories. The amino acid sequence at the HA cleavage site confirmed that the detected virus is a low pathogenic avian influenza virus (LPAIV), which is consistent with the absence of avian mortality or outbreaks in the Pantanal. Previous reports of LPAIV subtypes in other Brazilian states further support this classification. Genomic composition of H5N2 detected in the Pantanal suggests the possibility of reassortment events, potentially combining external gene segments related to Colombian lineages with internal segments derived from influenza viruses circulating in multiple regions (Table 3). This genomic composition highlights the role of migratory birds in facilitating long-distance viral gene flow and generating genetically diverse AIV strains in the Pantanal.

Our subtype may illustrate the potential for genetic reassortment commonly observed among LPAI viruses, although no formal reassortment analysis was conducted in this study. We also highlight the potential for inter-hemispheric viral dissemination mediated by migratory host species, reaffirming previous reports on the high genetic plasticity and reassortment capacity of H5 subtype AIVs. These characteristics favor evolutionary processes at migratory stopover and wintering sites, where the aggregation of individuals from distinct geographic regions increases the likelihood of interspecies viral exchange^30,36,37. Despite these ecological considerations, our genomic analysis did not identify any molecular markers associated with mammalian adaptation in the HA or PB2 segments, such as the PB2 E627K and D701N mutations, which have been widely linked to enhanced replication in mammalian hosts across multiple influenza A subtypes, including both HPAI and LPAI viruses as previously characterized in studies involving cross-species adaptation of AIVs^38,39. These findings suggest that no mutations associated with enhanced replication or host switching in mammalian species were identified, reinforcing the avian origin and host specificity of this H5N2 strain.

The detection of H5N2 LPAIV was initially performed in mixed swab samples (orotracheal and cloacal) from the shorebird Charadrius collaris, popularly known as the collared plover, captured in the Chacororé Bay, one of the largest lakes in the northern region of the Pantanal (Fig. 1). Subsequent analyses revealed different Ct values when orotracheal and cloacal swabs were tested separately, which may suggest a higher viral presence in respiratory secretions compared to intestinal samples. However, this interpretation should be approached with caution, as it is based on a single animal and the time of infection remains unknown. Nonetheless, the low viral detection in fecal samples collected from the environment aligns with the hypothesis of low viral shedding via the intestinal route, potentially consistent with the low pathogenicity observed for this subtype in the region.

The collared plover is a small, predominantly sedentary bird, but with records of short-range seasonal movements^40,41. In the Brazilian Pantanal, populations of C. collaris move within or outside the country during flooding in response to seasonal changes in water levels^27,42. Therefore, we assume that the infection of this bird occurred through environmental contamination in locations shared with migratory wild birds from Central and North America, especially from Mexico, where LPAI H5N2 subtype is enzootic^23,43,44. On the other hand, to unravel the origin of the ressortment event of this new H5N2 subtype, additional monitoring studies are necessary in locations where the presence of this bird species is recorded.

Covering approximately 8,515,000 km²⁴⁵ (IBGE), Brazil is the fifth largest country in the world by area, which highlights the critical need for continuous viral surveillance to ensure effective epidemiological monitoring. The first confirmed focus of H5N1 HPAIV infection in Brazil occurred in a Charadriiformes bird of the species Thalasseus acuflavidus from the State of Espirito Santo, located on the coast of southeastern Brazil⁴⁶. Since then, numerous animal health agencies in Brazil have worked tirelessly to prevent the spread of this subtype to domestic birds. Only now, two years after the first detection in T. acuflavidus bird, Brazil has the first focus of infection been reported on a breeding farm located in Rio Grande do Sul, in the extreme south of Brazil. Official measures to contain and eradicate the outbreak are in force to combat the disease and maintain the sector’s production capacity, guaranteeing food security for the Brazilian population⁴⁷.

We demonstrate for the first time a bird infected by LPAIV H5N2 subtype in the Brazilian Pantanal. Even in the face of a low pathogenicity virus, our result highlights the importance of epidemiological studies in avian populations in non-coastal areas that function as stopover sites for intercontinental migratory species thereby contributing to early detection efforts and global avian influenza preparedness.

Methods

Ethics approval and reporting standards

All methods were carried out in accordance with relevant guidelines and regulations, including those of the National Council for the Control of Animal Experimentation (CONCEA) of the Brazilian Ministry of Science, Technology and Innovation (MCTI). All experimental methods involving animals were approved by the Animal Use Ethics Committee of the Federal University of Mato Grosso (CEUA-UFMT, protocol number 23108.085496/2020-32) and authorized by the Brazilian Institute of the Environment and Renewable Natural Resources (IBAMA, permit number 10698-3).

All methods involving animals are reported in accordance with ARRIVE guidelines (https://arriveguidelines.org).

Studied region and sample collection

Owing to field constraints inherent to wild bird capture—including species occurrence, environmental unpredictability, and site accessibility—it was not possible to predetermine sample size using standard statistical methods^48,49. Consequently, sampling was driven by logistical feasibility to optimize temporal and species representation.

The field study was carried out between 2021 and 2023 in eight areas in the northern Pantanal region of the state of Mato Grosso, in the central-western region of Brazil. These sampling areas were selected prioritizing regions with a wide diversity of resident birds and a significant flow of migratory birds. At these points, we sought to sample different types of phytophysiognomy such as dry and flooded forests, capoes, riparian forests, beaches and flooded fields.

At each site, birds were captured using 15 mist nets (with a mesh size of 36 mm, 3.5 m in height and 10 m in length) that were set open from 6:00 am to 10:00 am for three to five consecutive days. Larger birds, such as waterfowl, were captured through active arrest with a net during the nighttime. Anatidae were captured through active arrest during the molting period (remiges period). The captured birds were marked with colored bands, and some species of interest were marked with metallic bands of the National Center of Research and Conservation of Wild Birds (CEMAVE).

After biometric measurements and species identification, oral and cloacal swabs were collected from each bird as follows; one mixed cryotube (combined orotracheal and cloacal swabs), one cryotube for the cloacal swab, and one cryotube for the oral swab. Swab samples were stored in cryotubes containing viral transport medium (VTM: 7.4 pH PBS + 10% glycerol + Antibiotics/fungizone)⁵⁰. After collection, samples were immediately frozen in liquid nitrogen. Upon arrival at the laboratory, swab samples were stored in freezers at –80° C.

Each specimen was identified taxonomically according to its order, family, scientific name, and common name. Subsequently, the birds were cataloged to facilitate classification according to their ecological niche, considering their relationship with the environment, habits, movement patterns, and diet. At the end of the entire procedure, the birds were released back into their natural habitat.

During the collection work, fresh feces were also collected from isolated flocks of water birds. The collection of these samples occurred after observing and identifying groups composed of a single species, to ensure that the material collected from the ground belonged to the set of birds being monitored. The fecal samples were placed in microtubes, labeled, and stored following the same protocol established for swab samples.

Molecular analysis

Samples of mixed swabs from the birds and fecal samples were subjected to molecular diagnosis for the detection of AIV. RNA was obtained using the MagMAX™ Core Acid Nucleic Purification Kit (Applied biosystems™, Austin, TX, USA) in an automatic DNA/RNA extractor MAGMAX™ FLEX (Applied Biosystems). After extraction, the genetic material was quantified in a NanoDrop™ (Thermo Fisher Scientific) and immediately analyzed.

Samples were then tested to identify the AIV using the AgPath-ID™ One-step RT-qPCR kit (Applied biosystems™, Austin, TX, USA) according to the manufacturer’s recommendations with primers and probe directed for the matrix gene^19,35. The amplification protocol began with reverse transcription at 45 °C for 10 min, primary denaturation at 95 °C for 5 min, followed by 40 cycles of denaturation at 94 °C for 10 s and annealing at 52 °C for 30 s. The amplification reactions were performed in QuantStudio™ real-time PCR thermocycler (Applied Biosystems™, Austin, TX, USA). A commercial equine influenza vaccine (Zoetis®, Lincoln, NE, USA) was used as the positive control, and ultrapure water was used as the negative control.

Sample presenting Cycle Threshold (Ct) below 38 for AIV were subjected to endpoint touchdown RT-PCR as described by Hoffman et al.⁵¹ and Zhou et al.⁵² to amplify the eight segments of AIV using the AgPath-ID™ One-Step RT-PCR Reagents (Applied biosystems™, Austin, TX, USA). The amplification protocol began with reverse transcription at 45 °C for 30 min followed by primary denaturation at 95 °C for 1 min, 5 cycles of 95 °C for 15 s, 47 °C for 15 s and 68 °C for 3 min and more 25 cycles of denaturation at 95 °C for 15 s, annealing at 57 °C for 15 s and extension at 3 min followed by a final extension at 68 °C for 7 min. The amplification reactions were performed in SimpliAmp thermocycler (Applied biosystems™, Austin, TX, USA). The amplification product was analyzed by electrophoresis on a 2% agarose gel stained with SyBR safe in E-Gel Power Snap Systems (Applied biosystems™, Austin, TX, USA). Products were visualized under UV light. RNA extracted from a commercial equine influenza vaccine was used as positive control.

The whole genome of Influenza A virus was obtained using Illumina® Microbial Amplicon Prep—Influenza A/B Kit (20106305) for the preparation of genomic libraries according to the manufacturer’s instructions from total RNA of the samples. The sequencing run was carried out in iSeq 100 Sequencing System (Illumina®) with 75 pM loading concentration as recommended for iSeq 100 platform, the run was set to generate 150 bp paired-end reads. Quality of the raw reads were checked with FastQC software and low quality reads were trimmed with Cutadapt⁵³. The Influenza virus segments sequences were recovered using SPADES software followed by taxonomic identification of obtained contigs with Kraken2^54,55. Classified sequences were validated as Influenza sequences with Blast using GenBank⁵⁶ database, higher identity sequences of each segment were used as reference for mapping the obtained reads using BWA⁵⁷ for coverage and completeness analysis (Table 3).Sequences alignment of obtained HA, NA, M1, NP, PA, PB2, NS and PB1 segments was carried out with MAFFT⁵⁸ using GenBank sequences.

After De Novo assembly, phylogenetic analyses were performed using an initial dataset of 1000 reference sequences obtained from GenBank. For enhanced resolution and analytical clarity, the dataset was curated by selecting the 500 most closely related sequences based on similarity metrics. The dataset was trimmed removing equal samples and low informative clades for better resolution. Phylogenetic analysis by Maximum Likelihood with IQ-Tree2⁵⁹ using GTR + F + I + G4 substitution model chosen by Model Finder⁶⁰. Bootstrap was set for 1000 for statistical support. Final trees were edited with iToL⁶¹.

Author contributions

Tayane Bruna Soares Magalhães: Conceptualization; Data curation; Investigation; Methodology; Visualization; Writing—original draft; Writing—review & editing. Elaine da Rosa Bueno: Data curation; Investigation. Nathalia de Assis Pereira: Investigation. Helena Aimée Santos Lima: Investigation. Sofia de Souza Pereira Gomes: Investigation. Priscila Juliana Sturmer Gonçalves Mizuuti: Investigation. Rachel Vieira Paes de Barros: Data curation; Formal analysis. João Batista de Pinho: Investigation. Thiago Borges Fernandes Semedo: Investigation. Edson Viana Massoli Junior: Data curation; Investigation. Luciano M. Thomazelli: Conceptualization; Writing—review & editing; Methodology. Ana Karolina Antunes Eisen: Investigation; Methodology. Nicole Almeida dos Reis: Investigation; Methodology. Desyrée Yumiko S. R. Ozaki: Investigation; Methodology. Giovana S. Caleiro: Investigation; Methodology. Gustavo O. Fenner: Investigation; Methodology. Fernanda Panicio Vizu: Investigation; Methodology. Theo Kraiser: Investigation; Methodology. Tatiana Ometto: Conceptualization; Writing—review & editing. Erick G. Dorlass: Data analysis; Writing – review & editing. Helena Lage Ferreira: Writing—review & editing. Clarice Weis Arns: Funding acquisition; Writing—review & editing. Edison Luiz Durigon: Funding acquisition; Resources; Validation; Writing—review & editing. Jansen de Araujo: Conceptualization; Formal analysis; Methodology; Resources; Supervision; Validation; Writing—review & editing. Daniel Moura de Aguiar: Conceptualization; Methodology; Resources; Supervision; Validation; Writing—review & editing. All authors reviewed and approved the final version of the manuscript.

Funding

This research was funded by the São Paulo Research Foundation (FAPESP; Grant/Award Number: 2023-03041-1); by the Brazilian Ministry of Science, Technology, and Innovation (MCTI- Brazil)/National Council for Scientific and Technological Development (CNPq; Grants/Award numbers 403761/2020-4 and 400172/2022-4; and by the Mato Grosso State Research Support Foundation (FAPEMAT; Grant/Award number 568520/2014) and Public Ministry Project Bank (MPMT; Grant/SIMP number 004172-096/2024). T.B.S.M., E.R.B, N.A.P., H.A.S.L., S.S.P.G., P.J.S.G.M., R.V.P.B., A.K.A.E., N.A.R., D.Y.S.R.O., G.S.C., G.O.F., F.P.V. and T.K. are recipients of Coordination for the Improvement of Higher Education Personnel (CAPES), CNPq and FAPESP scholarship. TBFS is supported by Fundação para a Ciência e Tecnologia (FCT), through PhD Grant/Award number 2022.10212.BD.

Data availability

The complete genome sequences originating from the H5N2 subtype detected in Pantanal are deposited in GenBank with accession numbers between PQ135836 and PQ135843. The full-length sequences of all eight genomic segments of the influenza A virus characterized in this study are publicly available under the following accession numbers: PB2: https://www.ncbi.nlm.nih.gov/nuccore/PQ135836. PB1: https://www.ncbi.nlm.nih.gov/nuccore/PQ135837. PA: https://www.ncbi.nlm.nih.gov/nuccore/PQ135838. HA: https://www.ncbi.nlm.nih.gov/nuccore/PQ135839. NP: https://www.ncbi.nlm.nih.gov/nuccore/PQ135840. NA: https://www.ncbi.nlm.nih.gov/nuccore/PQ135841. M: https://www.ncbi.nlm.nih.gov/nuccore/PQ135842. NS: https://www.ncbi.nlm.nih.gov/nuccore/PQ135843. Additional data are available from the corresponding author on reasonable request.

Declarations

Competing interests

The authors declare no competing interests.