Gene flow and its sporadic spillover: H10 and N5 avian influenza viruses from wild birds and the H10N5 human cases in China

Abstract

On January 30, 2024, China announced the first human case of H10N5 influenza infection. Prior to this, human cases of H10N7 and H10N8 had been reported. It is now appropriate to re-examine the evolution and future epidemiological trends of the H10 and N5 subtypes of avian influenza viruses (AIVs). In this study, we analyzed the reassortment characteristics of the first human-derived H10N5 AIV (A/Zhejiang/ZJU01/2023), as well as the evolutionary dynamics of the wild bird-derived H10 and N5 subtypes of AIVs over the past decade. Our findings indicate that the human-derived H10N5 AIV exhibited low pathogenicity. A/bean_goose/Korea/KNU-10/2022(H10N7) and A/mallard/Novosibirsk_region/962k/2018(H12N5) were identified as the potential reassortment parents. The virus has existed since 2022 and several isolations have been reported in Bangladesh. Phylogenetic analysis showed that H10Ny and HxN5 AIVs in China are clustered differently based on the East Asian-Australian (eastern) and Central Asian-Indian (western) migratory flyways. The H10Ny and HxN5 AIV reassortant strains may cause human infections through accidental spillover. It is possible that another center of AIV evolution, mutation, and reassortment may be developing along the migratory flyways in northeastern Asia, distinct from Europe, the Americas, and China's Yangtze River Delta and Pearl River Delta, which should be closely monitored to ensure the safety of the public.

Highlights

• •The epidemiological characteristics of the first human-derived H10N5 AIV were analyzed.
• •The potential reassortment parents of this human-derived H10N5 AIV were identified.
• •A clustering of AIVs in China was observed according to bird migratory flyways.
• •A new center for the evolution and transmission of AIVs is described.

Introduction

Influenza A viruses (IAV) belong to the family Orthomyxoviridae. Based on the antigenicity of the membrane glycoproteins hemagglutinin protein (HA) and neuraminidase protein (NA), they can be divided into 16 HA subtypes and 9 NA subtypes (Mostafa et al., 2018). With the exception of H17N10 and H18N11, which originated from bats, all IAV subtypes were initially isolated from poultry hosts, and almost all IAV subtypes have been classified as avian influenza viruses (AIVs) (Wang et al., 2022). AIVs can be divided into low pathogenicity (LP) and high pathogenicity (HP) strains according to their pathogenicity in chickens (Spackman, 2020). Wild waterfowl serve as natural hosts for AIV and play an important role in the evolution and spread of the virus. Although AIV is known to have a strong host species barrier, in recent years, a considerable number of AIVs have been isolated from animals other than birds.

A 63-year-old woman from Anhui Province, with a history of chronic comorbidities, has onset of symptoms on November 30, 2023. Experiencing with symptoms of fever, sore throat and cough, she was admitted to a nearby hospital on 2 December, transferred to a hospital in Zhejiang Province on 7 December, and died on 16 December due to the severe condition. Seasonal IAVs of H3N2 and H10N5 subtypes were isolated from the patient's specimens on 22 January (WHO, 2024). Although this is not the first case of human infection with H10 subtypes, it is the first case of human infection with H10N5 AIV reported globally. In fact, several H10Ny subtypes have been detected in wild birds and poultry worldwide since the discovery of H10N7 AIV in 1949 (Wu et al., 2019). H10Ny AIVs have also been successfully isolated from mammals, including minks (H10N4 and H10N7), seals (H10N7) and pigs (H10N5).

Viruses of the H10 subtype have also caused human spillover infections in recent years (Wang et al., 2012; Krog et al., 2015). In 2010, Australia reported several human cases of H10N7 infection with significant clinical symptoms (Arzey et al., 2012). In 2014, the first documented human case of H10N8 AIV infection occurred in China, which led to one fatality (Chen et al., 2014). H10 subtype AIVs are predominant strains circulating in poultry and wild birds, especially in live bird markets. Since most H10 subtype AIVs are LPAIVs, they typically do not exhibit clinical signs after infection in animals, making them difficult to detect (Pantin-Jackwood and Swayne, 2009). However, in the case of antigenic drift associated with virulence or pathogenicity, or antigenic shift resulting from reassortment with other subtypes, those viruses have the potential to cause sudden human epidemics.

In this study, we explored the genetic evolution and the potential reassortment parents of the first human H10N5 strain, as well as its phylogenetic characteristics. We also conducted a systematic analysis of the evolutionary dynamics and transmission routes of AIV subtypes H10 and N5 in China, aiming to establish a framework for future prevention and control of human epidemics caused by these types of AIVs.

Results

Glycosylation and cleavage sites

The ZJU01 isolate was found to have potential glycosylation sites in the HA and NA proteins that are consistent with those of other H10N5 strains. The HA sites 246, 262, and 392 were identified as N-X-T types, where X is S, R, and R, respectively. The glycosylation sites at positions 18, 135, and 182 were identified as N-X-S types, i.e., NMS, NHS, and NSS. There were two potential glycosylation sites of NA, 44 of which were N-X-T type as N-T-T, while 33 site was N-X-S type as N-I-S.

The cleavage site within the mature HA protein was identified at the amino acid position 334-PEVVQGR↓GLF-343 of the ZJU01 isolate. This exhibited characteristics consistent with a restriction site for Q/E-X-R LPAIV.

Comparison of internal genes

The ZJU01 isolate exhibited a high degree of nucleotide similarity in its internal genes PB1, PB2, PA, MP, NP, and NS when compared in NCBI to neighboring strains, including those from the Republic of Korea (R.O. Korea), the Democratic People's Republic of Korea (D.P.R. Korea), and Japan. The six strains exhibiting the highest nucleotide homology with ZJU01 are listed in Table 1. From the perspective of isolate subtype, the H3N8, H10N7 and H5 AIVs exhibited the highest nucleotide homology with ZJU01, and from the perspective of isolated hosts, the viruses are mainly found in wild aquatic birds, including wild ducks, bean geese, white-backed curlews, and white-fronted geese.

Table 1.

Comparison of A/Zhejiang/ZJU01/2023(H10N5) genes to reported strains with the highest identities.

Gene	Strains with the highest nucleotide homology	Subtypes	Reference Accession No.	Similarity
PB2	A/Anser albifrons/South Korea/22JN-163-1/2022	H10N7	OQ296821	98.73%
PB1	A/Eurasian_Curlew/China/CZ322(7)/2019	H3N8	MT835198	98.37%
PA	A/duck/Tottori/311215/2020	H5N2	LC656332	98.42%
NP	A/Wild duck/South Korea/KNU18-28/2018	H5N3	MT477770	97.80%
MP	A/environment/Japan/KU-4h/2021	H3N8	OR044153	99.39%
NS	A/Bean Goose(Anser fabalis)/Korea/KNU10/2022	H10N7	OR674059	99.77%

Phylogenetic analysis

The homology between the ZJU01 isolate and the H10 subtype AIVs was found to range from 74.2% to 97.2%. The lowest degree of homology was observed between the ZJU01 isolate and the A/Anser_fabalis/China/D211/2020(H10N8) strain, with an identity of 74.2%. Conversely, the highest degree of homology was observed between the ZJU01 and the A/Bean_Goose/Korea/KNU-10/2022(H10N7) strain, with an identity of 97.2%. Among the H10 subtype AIVs currently circulating in China, the ZJU01 isolate exhibited a high degree of homology with strains isolated from swans and Kentish plovers in the Shandong and Liaoning provinces. In particular, the ZJU01 isolate exhibited a similarity of ≥ 95.0% with the A/swan/Shandong/W4322/2020(H10N4) A/swan/Shandong/W4047/2020(H10N4), A/swan/Shandong/W3917/2020(H10N4), A/swan/Shandong/W3875/2020(H10N8), A/swan/Shandong/W4074/2020(H10N4), A/kentish_plover/Liaoning/DD646/2020(H10N9), A/kentish_plover/Liaoning/DD651/2020(H10N9) strain. The ZJU01 isolate is closely related to strains that have been identified in Bangladesh, R.O. Korea, and Japan. It belongs to the Eurasian clade, which also includes the Chinese strains H10N4, H10N8, and H10N9 that have been isolated in Shandong and Liaoning provinces. Furthermore, it was observed that the isolates from Liaoning were distributed across different clades, with some strains clustering with strains from Ningxia and Jiangxi provinces (Fig. 1A). The characteristics of the two clusters were reflected in the differences in amino acid sites, including 12S (A), 40D (N), 61I (L), 84S (A), 230 ​M (L), 278K (R), 285K (R), 309L (M), 320V (L), and 321V (M).

Fig. 1.

Phylogenetic analysis of H10Ny and HxN5 AIVs: geographic distribution and temporal separation patterns. A Phylogenetic tree H10Ny based on the hemagglutinin (HA) gene. B Phylogenetic tree HxN5 based on the neuraminidase (NA) gene. The red circle represents the first human isolate of H10N5 AIV, A/Zhejiang/ZJU01/2023(H10N5). The colored block representing the clade indicates the migration zone of migratory birds inferred from geographical information. The two color strips following the clades indicate the time and region of strain isolation.

A comparative analysis of the homology between the ZJU01 isolate and the N5 subtype of AIVs revealed a homology range of 92.2%–96.4%. The lowest homology observed was 92.2% with A/Anas_platyrhynchos/Belgium/8751_0001/2020(H12N5). In comparison, it shared the highest homology (96.0%) with the existing N5 subtype AIVs in China. The isolates were derived from wild birds in the Jiangxi and Hubei provinces, including A/Bean_Goose/Hubei/chenhu_XVI270-1/2016(H11N5), A/wild_bird/Jiangxi/01.17_JJCHBCT6/2019(H6N5), A/wild_bird/Jiangxi/01.17_JJCHBCT19/2019(H6N5), and A/wild_bird/Jiangxi/1.17_JJCHBCT20-2/2019(H6N5). Phylogenetic tree analysis revealed that the ZJU01 isolate clustered with strains from Bangladesh and D.P.R. Korea and was grouped together with the Russian epidemic strain in one phylogenetic cluster. Furthermore, strains from Jiangxi and Hubei provinces in eastern China also clustered with strains from Zhejiang province in two distinct branches (Fig. 1B), characterized by 39D (N), 76H (P), 89E (D), 185E (K), 196D (N), 366T (I).

Phylogenetic trees constructed from either the HA or NA gene exhibited the presence of distinct clusters of AIVs in China along the East Asian-Australian (eastern) and Central Asian-Indian (western) migratory flyways.

Selection pressure and population dynamics

The dN/dS of NP, PA, PB1 and PB2 was found to be significantly less than 1, indicating a strong negative selection pressure. Additionally, numerous purification selection sites were identified (Fig. 2). With regard to H10Ny, the HA, MP, NA segments were subject to strong positive selection pressure, while the NS segment also exhibited a certain degree of positive selection pressure (Fig. 2A). In contrast to H10Ny and H10N5, the MP segment of HxN5 was subjected to strong negative selection pressure, with no positive selection sites detected (Fig. 2B). With regard to H10N5, the dN/dS ratio of the HA, NA and NS segments exhibited a tendency towards 1 (Fig. 2C), indicating that the mutation in these segments were less affected by external factors. Conversely, the MP segment was subject to strong positive selection pressure.

Fig. 2.

Population dynamics and phylogenetic analysis of three AIVs populations. A, H10Ny. B, HxN5. C, H10N5. The bars show the selected sites confirmed by Single Likelihood Ancestor Counting (SLAC), Fixed Effects Likelihood (FEL), and Fast, UnconstrainedBayesian AppRoximation (FUBAR). The scatter points on the right indicates the value of dN/dS calculated based on SLAC, FEL, and FUBAR, with color indicating the direction of selection. The color of the circle indicates the predicted population dynamics results.

The results obtained from the DnaSP software indicated that the Tajima's D, Fu and Li's D and F values of the majority of genes of H10Ny and H10N5 were positive. In contrast, the Tajima's D, Fu and Li's D and F values of the majority of genes of HxN5 were negative. Consequently, it was postulated that the HxN5 population may be in an expansionary phase, whereas the H10N5 and H10Ny populations may be in a contractionary state.

Reassortment event of the human-derived H10N5 AIV

A total of 39 reassortment events related to the ZJU01 isolate were identified from 298 source strains, including 62 H10Ny and 45 HxN5 strains. The Simplot analysis showed that A/bean_goose/Korea/KNU-10/2022(H10N7) and A/mallard/Novosibirsk region/962k/2018(H12N5) were the possible parents of the ZJU01 isolate (Fig. 3A and B). Among them, A/bean_goose/Korea/KNU-10/2022(H10N7) provides gene segments of PB2, PB1, HA, MP and NS, while A/mallard/Novosibirsk_region/962k/2018(H12N5) provides PA, NP and NA segment for it. The sources of the six internal segments of the ZJU01 isolate are depicted in Fig. 3C and D.

Fig. 3.

Reassortment characteristics of the first human-derived H10N5 isolate. A Potential parental isolates of A/Zhejiang/ZJU01/2023(H10N5) identified from H10Ny. All of these isolates have been identified as parents by RDP based on their high scores. The red curve represents the parental strain identified by this study [A/bean_goose/Korea/KNU-10/2022(H10N7)], while the grey curves are the A/environment/Kagoshima/KU-J2/2021(H10N4), A/duck/Bangladesh/44491/2020(H10N7). A/mallard/Korea/KNU-2/2023(H10N4), A/kentish_plover/Liaoning/DD646/2020(H10N9), A/Eurasian_teal/South_Korea/JB32-15/2019(H10N7), A/Anser_albifrons/South_Korea163-1/2022(H10N7). B Potential parental isolates of A/Zhejiang/ZJU01/2023 (H10N5) identified from HxN5. The red curve is the parental strain we identified [A/mallard/Novosibirsk_region/962k/2018(H12N5)], and the grey curves are the A/wild_duck/shandong/W6808/2019(H12N5), A/mallard/Novosibirsk_region/3541k/2020(H12N5), A/Common_Teal/Chany_Lake/40/2019(H12N5), A/duck/Kyoto/261007/2014(H6N5), A/spot_billed_duck/Korea/KNU-17/2022(H6N5). C, D Simplot (C) and Bootscan (D) plots of the reassortment event of the A/Zhejiang/ZJU01/2023(H10N5). The blue curve represents the Major parent and the green curve represents the Minor parent.

In parallel, the reassortment events of H10Ny and HxN5 AIVs in China were summarized according to the calculation results of the RDP software (Fig. 4). It was found that wild aquatic birds were the main trans-regional carriers of the reassortant, with the majority entering China from neighboring countries via the wild bird flyway (mainly the East Asian-Australasian migratory flyway, and followed by the Central Asian-Indian migratory flyway), particularly in the eastern coastal areas.

Fig. 4.

The reassortment events in China based on the identification of RDP software. The first column is the host reservoir from which the reassortant isolate was detected, the second column is the region to which the reassortant parent belongs, the third column is the inferred route of influx of the reassortant isolate, and the fourth column is the region in which the reassortant isolate was detected in China.

Gene flows of the human-derived H10N5 AIV

Kst and Fst are measures of the relative proportion of total genetic diversity between populations, with values ranging from 0 to 1. The closer the values of Kst and Fst to 1, the greater the genetic differentiation between populations. Conversely, the higher the similarity between populations, the smaller the genetic differentiation. The Nm value can be used to measure the frequency of gene flow between populations. The higher the value, the more frequent the gene flow.

Our data indicated that H10N5 strains in China exhibited the highest similarity with those in R.O. Korea, D.P.R. Korea, Mongolia, Russia, Bangladesh, India and other neighbors, accompanied by a high frequency of gene flow between populations. In contrast, the genetic differentiation with the American epidemic isolate was relatively high, accompanied by a low frequency of gene flow (Fig. 5).

Fig. 5.

Gene flow and degree of genetic differentiation of H10N5 subtype AIVs in China with neighboring countries. A, HA gene. B, NA gene. The x-axis represents Fst values, which measure the genetic differentiation among strains from different regions; higher values indicate a more significant genetic structural difference between populations. The y-axis represents Nm values, which assess gene flow among strains from different regions; higher values suggest more active gene flow and frequent genetic exchange between populations. The three colors represent comparisons of population differences between strains isolated in China and those from neighboring countries in Southwest Asia (India, Pakistan, Bangladesh, etc.), Northeast Asia (Japan, Russia, D.P.R. Korea, R.O. Korea, etc.), and North America (primarily the United States).

Discussion

Since the first isolation of H10 subtype AIV from chickens in Germany, H10Ny isolates have been frequently detected in poultry and wild waterfowl worldwide, and a number of novel reassortant human cases of H10 subtype AIVs have been reported in recent years (Arzey et al., 2012; Kim et al., 2012; Lee et al., 2021). Studies have shown that H10Ny strains have become more adapted to mammals (Lv et al., 2023). In our study, the initial human-derived H10N5 isolate ZJU01 may be a reassortant of H10N7 and H12N5 subtypes of AIVs. The strain of A/bean_goose/Korea/KNU-10/2022(H10N7) provided it with five genomic segments, PB2, PB1, HA, MP, and NS, while the strain of A/mallard/Novosibirsk_region/962k/2018(H12N5) provided it with PA, NP, and NA segments. The similar strains had already appeared in 2022, with several isolates having been recorded in Bangladesh (Fig. 6A).

Fig. 6.

Spatial and temporal dynamics of the novel H10N5 AIV origins and global reassortment events. A Spatial and temporal model of the origins of the novel H10N5 AIV. B Bivariate heat map of East Asia based on AIV influx and outflow. Inflow degree measures the number of reassorted strains in the region, while outflow degree quantifies the frequency with which strains from the region serve as parental strains. C Network of global AIV reassortment events. Each line represents a reassortment event. The three flags mark the locations where the first global human H10N5 isolate was detected and the regions to which the major and minor parental strains belong.

Phylogenetic analyses confirmed the above conclusion. The phylogenetic trees of the HA and NA genes indicated that ZJU01 could be clustered into isolates originating from the East Asian-Australasian migratory flyway and the Central Asian-Indian migratory flyway, respectively. It is noteworthy that D.P.R. Korea, Russia and Bangladesh are also located within these two migratory flyways (Nam et al., 2021). The eight segments of ZJU01 showed the highest identity to AIVs isolated from countries along the East Asian-Australian migratory flyway, including Japan, R.O. Korea, and D.P.R. Korea. A number of reassortant H10Ny and HxN5 isolates were also found in Jiangxi, Jiangsu, Hubei, Shandong, Liaoning and other provinces in China. Isolates from Shandong, Jiangsu and Liaoning are geographically close to those from Jiangxi and Hubei, but still within different phylogenetic branches. This complex distribution suggests that the migratory flyways of wild waterfowl along the northeast Asian coastline of China may become another center for the evolution and reassortment of AIVs, distinct from Europe, the Americas, and the Yangtze and Pearl River deltas of China. It is recommended that this area should be closely monitored (Fig. 6B and C).

AIV has spread globally due to its high mutability, transmissibility and the ability of its natural hosts (birds) to migrate long distances (Health et al., 2017). The seasonal migration and mass aggregation behavior of birds has been shown to coincide in time and space with the epidemic trend of AIV influenza (Shi et al., 2018). Due to the vast area of China, three major wild bird migratory flyways in the Asia-Pacific region pass through the country, with tens of millions of birds wintering there (Gu et al., 2021). As a result, a variety of different subtypes of AIVs are transported across different areas, often reassorting with prevalent strains in neighboring countries and showing no significant genetic differentiation (Lin et al., 2023). Gene flow analysis has shown that there is significant genetic exchange of epidemic AIVs between China and several other countries, including Russia, R.O. Korea, D.P.R. Korea, Japan, Bangladesh, Egypt, Vietnam, and Mongolia. However, given that the eastern coast of China lies within the East Asian-Australian migratory flyway, the western interior lies within the Central Asian-Indian migratory flyway, and the vast central region of the country lies at the junction of these two migratory flyways, the driving force of viral strain reassortment varies between provinces. The prevalent virus strains in the eastern areas, represented by Jiangsu, Jiangxi, and Zhejiang, primarily reassort with strains carried by birds from R.O. Korea, D.P.R. Korea, Japan, and Vietnam, while those in the western areas, represented by Chongqing, Sichuan, and Ningxia, primarily reassort with strains carried by birds from Egypt and Bangladesh. However, the migration of wild birds is influenced by a variety of factors and does not follow strict migratory flyways, which has led to the detection of virus strains carried by birds from Bangladesh and Egypt in Zhejiang and Jiangsu.

In terms of avian hosts, transcontinental transmission was facilitated primarily by wading birds, represented by Anseriformes and Charadriiformes. This finding is in line with previous studies (Bevins et al., 2014). The main species were ducks, including Anas crecca, Anas platyrhynchos, Anser albifrons, Anas clypeata, Cygnus olor, Anser anser, and Anas zonorhyncha. A study has shown that ducks are more likely to be hosts for AIVs than geese or chickens (Lin et al., 2023). A commonly accepted mechanism for the environmental transmission of AIVs is the excretion of faeces by birds, which contaminates water sources and environmental surfaces, and other animals could be infected by oral ingestion of contaminated water, soil or faeces (Fourment and Holmes, 2015). The species of Anas crecca and Anas platyrhynchos are particularly abundant and tend to inhabit shallow water habitats. Some studies have reported high levels of AIVs in the aquatic environment of their habitats (Hill et al., 2010; Hill and Runstadler, 2016). Combined with the complexity and high density of wild duck species in China, the high prevalence of AIV infection in duck family broods, and the persistence of AIVs in water (Nazir et al., 2010), it can be broadly concluded that the reassortment of AIVs in China is characterized by multi-subtype reassortment in different areas along the East Asian-Australian migratory flyway in the east and the Central Asian-Indian migratory flyway in the west, and in different species of wild bird hosts, which also illustrates the importance of wild birds in the transmission and genetic evolution of AIVs.

The selection pressure on the virus may reflect the immune status of the host and the pattern of viral evolution (Haasdijk and Heinerman, 2018). AIV, like other RNA viruses, forms quasispecies in infected hosts due to the lack of corrective function of the RNA polymerase (Dietze et al., 2018). Once under pressure from the host immune system, those quasispecies that favor immune escape are selected for and strengthened (Fourment and Holmes, 2015). Indeed, some AIV strains have been characterized by stable antigenic drift since their emergence (Cattoli et al., 2011). In this study, we found that the H10Ny, HxN5 and H10N5 populations showed strong purifying selection on PB1, PB2, PA and NP. The common components of RNA polymerase in AIV, PB1, PB2 and PA, are also key obstacles to its adaptation to a new host across the species barrier (Nilsson et al., 2017). NP also plays an important role in AIV replication and host adaptation. Strong negative selection pressure and no significant variation were observed at positions 591, 627 and 701 of PB2, positions 216 and 524 of PB1, positions 97 and 522 of NP and position 319 of NP in the three AIV subtypes. This suggests that there is no significant breakthrough in host tropism between the three AIV subtypes. MP encodes an ion channel protein, while NS inhibits the host antiviral response by suppressing interferon and other antiviral proteins. Both proteins play a key role in AIVs invasion, assembly and release (Tindale et al., 2020; Kim et al., 2021). Meanwhile, mutations in MP can increase resistance to amantadine. We found mutations at positions 30 and 31 of the M2 protein in some H10Ny strains, but no similar phenomenon was observed in ZJU01 strains or HxN5 populations. The HA protein can interact with sialic acid and is the main mediator of viral adhesion and invasion. The NA protein can cleave sialidase and mediate the release of progeny virus from cells. Some studies have suggested that there is a functional balance between HA and NA, which together influence the adaptability of AIV to the host (Lakdawala et al., 2015). The higher selection pressure on MP and HA in the H10Ny population has resulted in greater functional adaptability and associated drug resistance mutations, suggesting that H10Ny should be more vigilant than HxN5 in causing severe clinical symptoms in humans.

Our study found that there were amino acid polymorphisms in HA and MP of the ZJU01, which were positive selection sites, suggesting that the novel reassortant persistently infects the host by altering some amino acids of HA and NA. Despite the presence of positive selection sites, most amino acid sites in ZJU01 tended to be under purifying selection, suggesting that the newly generated virus ZJU01 had eliminated deleterious site mutations after crossing the species barrier (from birds to humans) and produced more functionally conserved proteins to adapt to the body's immune system. In addition, population dynamic analysis suggested that the H10N5 and H10Ny AIVs populations showed a contraction trend. This indicates that the H10N5 human AIV strain that emerged this time was an accidental spillover infection in humans, and the possibility of it causing a large-scale epidemic in the future is extremely low. In the studies of Tajima's D test and Fu and Li's test, we employed sequences spanning a longer temporal range to include more genetic variation, allowing for a more accurate assessment of the dynamic changes and evolutionary processes in the populations under investigation, as well as to enhance the statistical power of the tests. However, this approach does have certain inherent limitations, such as the potential complexity in interpretation due to specific environmental events or selective pressures occurring at different times. Despite these limitations, the results remain valuable, particularly given the currently limited number of available strains, especially H10N5 strains. These results underscore the need to pay closer attention to the public health risks posed by N5 subtype strains during epidemic prevention and control, while also striving to comprehensively collect genomic data of avian influenza viruses from both the environment and hosts.

Conclusions

In conclusion, novel reassortants of H10Ny and HxN5 AIVs have been demonstrated to infect mammals, including humans, which may cause public health issues. However, they are unlikely to develop an epidemic trend in the short term. In addition to surveillance of H5, H7 and H9 subtypes, continuous surveillance and biosafety assessment of other AIV subtypes and their evolution are important. Our study confirmed that ZJU01 was an accidental spillover and identified its parents, providing an early warning of the potential risks of H10Ny and HxN5. Despite the lack of large samples, it suggests that there are significant differences in AIV reassortment and prevalence between eastern and western China, particularly along the flyways of northeast Asian migratory birds in China. This area is gradually becoming another center of AIV evolution, variation and reassortment, distinct from Europe, America and the Yangtze and Pearl River deltas in China, and should be closely monitored.

Materials and methods

Sequences and genomes preparation

The genomes of the H10Ny and HxN5 AIVs were obtained from two databases: the Global Initiative on Sharing Avian Influenza Data (GISAID, https://gisaid.org/) and the National Center for Biotechnology Information (NCBI, https://www.ncbi.nlm.nih.gov/). In order to analyze the original human H10N5 strain (A/Zhejiang/ZJU01/2023(H10N5), abbreviated as ZJU01), the Basic Local Alignment Search Tool (BLAST, https://blast.ncbi.nlm.nih.gov/) on the NCBI website was employed to identify the 100 sequences most closely related to each of its internal protein genes. The genomes of all the isolates, approximately 600, were then extracted and the duplicates were removed. All sequences used to conduct phylogenetic and evolutionary analyses were listed in Supplementary Table S1.

Glycosylation and cleavage sites prediction

To determine the biological characteristics of the hemagglutinin (HA) and neuraminidase (NA) genes, we conducted an analysis of the amino acid sequences derived from the nucleotide sequences of the ZJU01 isolate and H10N5 strains obtained from the aforementioned databases. Subsequently, the online tool NetNGlyc was used to predict glycosylation sites (https://services.healthtech.dtu.dk/services/NetNGlyc-1.0/). Additionally, we predicted and analyzed the cleavage sites of the ZJU01 isolate by referencing the HA genes of multiple H10 subtype AIVs from the NCBI nucleotide database.

Phylogenetic analysis

The ClustalW program in MEGA11 software was employed for multiple alignments, and the Find Best-Fit Substitution Model function was utilized to calculate the optimal tree-building model. Phylogenetic trees for the HA and NA genes were constructed using the neighbor-joining method, the TN93 model, and a gamma distribution, respectively, with a bootstrap value of 1000. Subsequently, the phylogenetic trees were modified using the online platform iTOL (https://itol.embl.de/tree/). The BioEdit 7.2.5 software was employed to analyze the homology of the eight genomic segments.

Selection pressure and group dynamics analysis

All stop codons were removed. Selection pressure on the HA and NA genes was determined using the Fixed Effects Likelihood (FEL), Fast, UnconstrainedBayesian AppRoximation (FUBAR), and Single Likelihood Ancestor Counting (SLAC) methods, respectively, using the online analysis website Datamonkey (http://www.datamonkey.org/). Synonymous substitution rate (dS), non-synonymous substitution rate (dN), Tajima's D value, Fu and Li's D and F values were calculated using DnaSP6·12·03 software. The selection mode was determined based on the dN/dS value. If dN/dS ​> ​1, it indicates positive selection. If 0 ​< ​dN/dS ​< ​1, it indicates purifying selection. The subtype dynamics of AIV were assessed using Tajima's D value, Fu and Li's D and F values. Tajima's D test identified evolutionary events such as population expansion, bottlenecks and selection by comparing the estimated number of segregating sites with the mean pairwise difference between sequences. The Fu and Li's D and F tests are particularly sensitive to population demographic expansion, with a negative value typically observed in expanding populations.

Reassortment event detection

The genomic segments of each H10Ny and HxN5 isolate, as well as those derived by BLAST of the six internal protein genes within the ZJU01 isolate, were aligned using MAFFT. The segments belonging to a specific isolate were spliced in sequence according to the segment number. The genome FASTA matrices were assembled based on AIV subtypes, resulting in the following configurations: PB2-PB1-PA-HA-NP-MP-NS (H10Ny), PB2-PB1-PA-NP-NA-MP-NS (HxN5), and PB2-PB1-PA-NP-MP-NS (HxNy, which includes H10Ny and HxN5). In each matrix, the genome of the ZJU01 isolate was designated as the subject of the query. Six methods (RDP, GENECONV, MaxChi, Chimaera, SiScan, and 3Seq) were employed using the RDP4 software package to identify potential reassortment events within the H10Ny, HxN5, and HxNy population. The identified reassortment event would then be subjected to further validation in the Simplot3·5·1 software.

Population genetic differentiation, and gene flow

The extent of genetic differentiation among populations was evaluated by Fst, and the significance of this differentiation was examined by two permutation-based statistical tests, Kst and Snn. The number of migrants successfully entering a population per generation (Nm) was employed as a metric for quantifying gene flow between populations. All of these values were calculated using the DnaSP6·12·03 software.

Data availability

All the data generated during the current study are included in the manuscript. Additional datasets and materials are available upon reasonable request from the corresponding author.

Ethics statement

This article does not contain any studies with human or animal subjects performed by any of the authors.

Author contributions

Weijie Chen, Shuiping Lu, Haiyan Xiong, and Chenglong Xiong were responsible for writing-original draft; Weijie Chen, Zhuiyu Xiang, Yuxi Wang, and Jingjing Hu conducted the investigation; Yue Pan, Yanjiao Li, Qile Gao, Siru Hu, and Qi Chen performed data curation and formal analysis; Haiyan Xiong, Chenglong Xiong, and Weibing Wang conducted validation and supervision; Chenglong Xiong and Weibing Wang took on supervision. All authors contributed to writing-review & editing, and all authors have seen and approved the final version.

Conflict of interest

All authors declare no competing interests.

Appendix A. Supplementary data

The following are the Supplementary data to this article: