|Home | About | Journals | Submit | Contact Us | Français|
Multiple reassortment events between different subtypes of endemic avian influenza viruses have increased the genomic diversity of influenza viruses circulating in poultry in southern China. Gene exchange from the natural gene pool to poultry has contributed to this increase in genetic diversity. However, the role of domestic ducks as an interface between the natural gene pool and terrestrial poultry in the influenza virus ecosystem has not been fully characterized. Here we phylogenetically and antigenically analyzed 170 H6 viruses isolated from domestic ducks from 2000 to 2005 in southern China, which contains the largest population of domestic ducks in the world. Three distinct hemagglutinin lineages were identified. Group I contained the majority of isolates with a single internal gene complex and was endemic in domestic ducks in Guangdong from the late 1990s onward. Group II was derived from reassortment events in which the surface genes of group I viruses were replaced with novel H6 and N2 genes. Group III represented H6 viruses that undergo frequent reassortment with multiple virus subtypes from the natural gene pool. Surprisingly, H6 viruses endemic in domestic ducks and terrestrial poultry seldom reassort, but gene exchanges between viruses from domestic ducks and migratory ducks occurred throughout the surveillance period. These findings suggest that domestic ducks in southern China mediate the interaction of viruses between different gene pools and facilitate the generation of novel influenza virus variants circulating in poultry.
Southern China has long been considered a hypothetical epicenter for facilitating the emergence of pandemic influenza viruses (13, 14). Systematic influenza surveillance has shown that the ecology of influenza virus in southern China has become increasingly complex in the last 2 decades (18). Multiple subtypes of influenza viruses have emerged, causing repeated poultry outbreaks, human infection, and dissemination to many countries in Eurasia that has given rise to a persistent pandemic threat. Epidemiological and genetic analyses revealed that three subtypes of influenza viruses, including highly pathogenic avian influenza (HPAI) H5N1 virus and low-pathogenic H9N2 and H6N1 subtypes, cocirculated within a variety of poultry hosts in this region. These viruses have also undergone extensive reassortment either between different endemic virus subtypes or with viruses from the influenza virus natural reservoir (3, 5, 23).
In the last decade, major research efforts have focused on understanding the emergence and development of the above viruses (1, 3, 8, 17, 22, 23). As approximately 60% of the world population of domestic ducks (approximately 600 million birds) are farmed in southern China (7), this represents the largest aquatic bird reservoir for influenza viruses in the world. Yet many subtypes of influenza viruses isolated from this population have not been systematically studied. Thus, the mechanisms for virus emergence and interaction in this population remain to be determined.
Aquatic birds are known to host a high diversity of avian influenza A viruses (AIV) (6, 19). All domestic ducks, except Muscovy ducks, are derived from mallards, and these hosts represent an interface between the aquatic bird natural AIV reservoir and the domestic poultry population (1, 17). Regular transmission of AIV between the aquatic reservoir and the domestic birds in China has resulted in increased genetic diversity of all sublineages known to circulate in domestic birds (5, 17). However, the movement and interaction of influenza virus between migratory and domestic ducks have not been fully explored. As H6 subtype viruses are among the most frequently encountered in our surveillance, in both wild and domestic aquatic birds, we have analyzed the evolutionary genomics of this subtype to elucidate the role that domestic ducks play in the influenza virus ecology of this region.
Genetic characterization of 170 H6 subtype representative viruses isolated from 2000 to 2005 in southern China revealed genetically distinct H6N2 hemagglutinin (HA) lineages in domestic ducks. Our results show that these viruses reassort much less frequently than do other influenza A viruses, i.e., they maintain a stable genome constellation through time and space and have been endemic in domestic ducks in Guangdong since the late 1990s. Our findings also showed that the replacement of the surface genes occurred and resulted in the establishment of a new H6N2 virus lineage that largely maintained the same internal gene complex. The epidemiological and ecological behavior of H6 viruses in domestic ducks highlights properties of the influenza virus ecosystem in southern China that have not been observed before. In particular, our results show for the first time that a unique nonreassortant lineage has become endemic in domestic ducks in the region. Our findings also demonstrate that gene exchange between the domestic duck and migratory duck gene pools occurs frequently but that there is only limited gene flow between H6 viruses endemic in terrestrial hosts and domestic ducks. Furthermore, we found that the sources for some gene segments currently circulating in the dominant H5N1 variants were likely H6 viruses introduced from migratory ducks. The findings of our study suggest that domestic ducks in southern China have a dual role in the influenza virus ecosystem—as a component of the natural reservoir along with wild aquatic birds but also as a major poultry type that harbors endemic virus lineages.
Avian influenza virus surveillance was conducted in live-poultry markets in six provinces of southern China from July 2000 to December 2005. Cloacal and tracheal swabs were taken from apparently healthy poultry if possible; otherwise, fecal swabs were collected. Sampling was conducted weekly. The surveillance program was initiated in Guangdong province (Shantou) from July 2000 and then extended to Hunan and Jiangxi in 2002, followed by Guangxi and Fujian in 2004 and Guizhou in 2005. In Jiangxi, sampling was carried out on sentinel duck farms located on Poyang Lake and from 2004 on was extended to capture migratory ducks at the lake (2). In all other sites, poultry types sampled include chicken, goose, duck, and several minor poultry species. Occasionally mallard-like birds that could be distinguished from domestic ducks at the live-poultry markets were labeled as wild duck to indicate this distinction. It is unclear whether these mallard-like birds were bred in captivity or were captured migratory ducks. Virus isolation and subtype identification were conducted using standard protocols as previously described (11). One hundred seventy H6 viruses were randomly and proportionally selected from the positive candidates based on surveillance region, sampling occasion, and isolation rate as representatives for further analyses.
Antigenic characteristics of the representative H6 influenza viruses were compared by hemagglutination inhibition (HI) assay with postinfection chicken antisera raised against chicken/Hong Kong/17/1977 (Ck/HK/17/77, H6N1), laughing gull/Delaware/4/1994 (LG/DE/4/94, H6N8), teal/Hong Kong/W312/1997 (Te/HK/W312/97, H6N1), quail/Hong Kong/1721-20/1999 (Qa/HK/1721-20/99, H6N1), quail/Hong Kong/YU1564/2000 (Qa/HK/YU1564/00, H6N?), and duck/Jiangxi/227/2003 (Dk/JX/227/03, H6N1) that were generated in our laboratory.
RNA extraction, cDNA synthesis, and PCR were carried out as previously described (21). Sequencing was performed for representative viruses by using a BigDye Terminator v3.1 cycle sequencing kit on an ABI 3730 genetic analyzer (Applied Biosystems) according to the manufacturer's instructions. DNA sequences were compiled and edited by using Lasergene 8.0 (DNAstar, Madison, WI).
Multiple sequence alignment of representative H6 viruses was performed together with alignment of sequences downloaded from GenBank using BioEdit 7.0 (9). Full-length gene sequences for each segment from the first start codon were used for phylogenetic analyses. Phylogenetic analysis was based on the following nucleotides: PB2, 1 to 2280; PB1, 1 to 2274; PA, 1 to 2151; HA, 1 to 1704; NP, 1 to 1494; NA, 1 to 1410; M, 1 to 982; and NS, 1 to 844. The program MrModeltest 2.2 was used to determine the appropriate DNA substitution model and rate heterogeneity (12). The generated model was used in all subsequent analyses. Neighbor-joining and maximum-likelihood trees were constructed using PAUP* 4b10 (16) and Garli 0.96, respectively (24). Bayesian Markov chain Monte Carlo (MCMC) analysis of two independent replicates of 2 million generations with six chains and sampling every 200 generations was conducted with MrBayes 3.1 (10). Estimates of the phylogenies were calculated by performing 1,000 neighbor-joining bootstrap replicates, and Bayesian posterior probabilities were calculated from the consensus of 18,000 trees after exclusion of the first 2,000 trees as burn-in.
The nucleotide sequences obtained in the present study are available from GenBank under accession numbers HM144388 to HM145740.
A total of 153,063 swabs were collected from apparently healthy poultry in southern China from 2000 to 2005. Sixteen percent of the total influenza virus isolates (1,325/8,325) were of H6 subtype with 413 isolated from minor poultry (partridges, quails, etc.) and 1 from chicken. The remaining majority were isolated from domestic ducks and geese (911/1,325). Isolation rates in winter and spring were generally higher than those in summer and autumn, although H6 viruses could be identified year-round. H6N2 was the most prevalent subtype among all H6 viruses from aquatic birds, followed by H6N1 and others, such as H6N5 or H6N6. We found only 2% positive isolates from tracheal swabs in aquatic poultry (16/911), and the remaining 98% of viruses were isolated from cloacal or fecal swabs, suggesting that those H6 viruses replicate primarily in the gut.
A total of 170 H6 viruses (about 20%) isolated from aquatic birds (ducks, n = 165; geese, n = 5) were selected to represent positive sampling occasions for sequence analysis. At least one virus was sequenced from each positive sampling occasion. For those samples having more than two isolates, viruses were proportionally sequenced.
Phylogenetic analysis of the H6 hemagglutinin (HA) gene classified the duck isolates into three major lineages or clades comprising isolates from both wild and domestic aquatic birds. The first group (group I), represented by Dk/ST/339/00 (ST339-like), comprised those viruses isolated exclusively from Guangdong (Shantou) and Hong Kong. This lineage formed a monophyletic group with a common ancestor isolated in Hong Kong in 1998 (Dk/HK/1037-1/98) (Fig. (Fig.1a).1a). Within this group the majority of the isolates were H6N2 (n = 81) or 51% of total H6 viruses tested. The other neuraminidase (NA) subtype combinations were much less frequently encountered, and no wild bird or sentinel duck isolates phylogenetically grouped with these taxa. This indicated that these viruses represented a stable H6 lineage in domestic ducks from this region.
The second group (group II), represented by WDk/ST/2853/03 (ST2853-like), had both H6N2 and H6N6 viruses that were isolated from Guangdong and Fujian (Fig. (Fig.1a).1a). This clade circulated in domestic ducks from 2002 to 2005 and cocirculated with group I in at least Guangdong province. Thirty-two percent of the H6 isolates sequenced clustered into this clade. This clade had a sister relationship with those H6 subtype viruses, represented by teal/HK/W312/97 (W312-like), established in terrestrial minor poultry in southern China since at least 1997 (4). Only two isolates from this study grouped with previously described H6 viruses from terrestrial poultry (WDk/ST/867/2002 and Dk/ST/1275/2004), suggesting that those established terrestrial bird viruses were rarely transmitted back to aquatic birds.
The third group (group III), represented by Dk/HN/573/02 (or HN573-like), was composed of those viruses isolated from inland regions such as Jiangxi and Hunan (Fig. (Fig.1a),1a), and its members were phylogenetically related to H6 viruses isolated from wild or migratory birds and included some viruses isolated from poultry disease outbreaks from different countries (Fig. (Fig.1a).1a). Viruses of this clade obviously belong to the Eurasian gene pool and included the biggest diversity of NA subtype combinations, including N1, N2, N5, and N8 (Fig. (Fig.1b).1b). Most importantly, all isolates from the sentinel duck farm in Jiangxi and from domestic ducks in Guangxi and Hunan, but only a single isolate from Shantou, belong to this group. This suggests that within Guangxi, Hunan, and Jiangxi there is a greater degree of interaction between migratory and domestic ducks than there is for lineages established in Guangdong province. It has been noted that those H6 viruses introduced into North America (represented by Mall/MN/Sg-00776/08 H6N2) also belong to this lineage.
The majority of the H6 subtypes encountered in our surveillance were H6N2 (79%). Other H6 subtypes, including H6N1 (8 isolates), were rarely encountered. Phylogenetic analysis of the N2-NA gene tree showed that it could be delineated into three major lineages (group I, group II, and group III) (Fig. (Fig.1b).1b). Group I formed a strongly supported monophyletic clade primarily comprised of aquatic isolates from Shantou (ST339-like). However, two isolates from Hunan (Dk/HN/908/2005 and Dk/HN/989/2005) were also nested within this clade. Similarly to the HA phylogeny, this clade was not associated with any wild aquatic bird and sentinel duck isolates, a finding that confirmed a distinct lineage unique to domestic ducks. Group II was closely affiliated with aquatic bird isolates from neighboring countries and with different HA subtype combination. Isolates of group III (HN573-like) included those from sentinel and migratory ducks in Jiangxi and domestic ducks from Hunan and Guangxi as well as those from Eurasian countries and represent a natural gene pool and transmission between the wild reservoir and domestic ducks. Only one N2 gene from WDk/ST/867/02 clustered with terrestrial H6/H9N2-like viruses that have become endemic in this region since the early 1990s (Fig. (Fig.1b1b).
In general, for all internal gene trees the majority of aquatic source viruses formed a monophyletic group identified as group I, which included the ST339-like viruses (Fig. (Fig.22 and and3).3). Also, in all internal gene trees, gene segments circulating in the natural reservoir could also be identified and classified as group III. However, unlike the surface proteins, group II viruses rarely formed a monophyletic clade in the internal gene phylogenies. In some clades H6 viruses clustered with HPAI H5N1 reassortant viruses. However, in all trees terrestrial and aquatic H6 isolates were distinct and rarely mixed.
Phylogenetic analyses of the PB2 gene demonstrated that six lineages cocirculated in Eurasia. H6 isolates tested were incorporated into each of those lineages (Fig. (Fig.2a).2a). The majority of H6 viruses from aquatic sources were monophyletic (group I) and formed a sister relationship with W312-like terrestrial H6 viruses. It was noted that the representative viruses for group I and group II were all clustered together in this lineage. The remaining H6 viruses joined two clades that contained both wild and domestic duck isolates and belonged to the natural gene pool corresponding to those group III viruses in the HA phylogenetic tree. The second clade comprised a mix of domestic and wild birds, including H5N2 isolated from migratory ducks (MDk/HK/MP206/05) in Hong Kong. Notably, two H6 isolates grouped in this clade, with Dk/HN/491/05 falling into the basal portion of the lineage leading to the dominant H5N1 variants (represented by Ck/HK/YU22/02 [genotype Z] and MD/JX/2136/05 [genotype V]) (5, 17). The remaining lineages belonged to those H5N1 or H9N2 lineages established in this region during the past 2 decades.
Similar phylogenies were observed for the PB1 and PA genes. In the PB1 gene tree, the majority of characterized H6 viruses clustered together and formed the group I lineage, while viruses isolated from Jiangxi, Hunan, and Guangxi joined the Eurasian gene pool lineage. It has been noted that about 10 H6 viruses isolated from wild duck (WDk/ST/2395/01) in Shantou formed another independent lineage with a sister-group relationship to the Gs/Guangdong/1/96-like highly pathogenic H5N1 virus lineage (Fig. (Fig.2b).2b). None of the aquatic lineages formed a strong phylogenetic relationship with terrestrially derived H6 viruses. The group I virus lineage had a sister relationship with Ck/Beijing/1/94-like H9N2 virus lineage with strong bootstrap and Bayesian support, suggesting that these two lineages might be derived from a common ancestor recently (Fig. (Fig.2b).2b). Phylogenetic analysis of the PA gene segment showed that group I viruses were sisters to established terrestrial H6N1 (W312-like) and H9N2 (G1-like, CK/Bei-like) viruses (Fig. (Fig.2c).2c). The remaining H6 viruses incorporated into different lineages of the Eurasian influenza virus gene pool.
Phylogenetic analyses of the NP, M, and NS genes all confirmed the establishment of an H6 virus lineage in domestic ducks in southern China which occupied the main body of viruses tested and covered the whole sampling period. Viruses of this lineage were detected from birds in Shantou and Fujian. H6 viruses from Hunan, Guangxi, and Jiangxi, isolated from either domestic, sentinel, or migratory ducks, were clustered into different lineages that belonged to the Eurasian influenza virus gene pool, suggesting multiple transmissions or gene mixing between migratory waterfowl and domestic aquatic birds (Fig. (Fig.3).3). Interestingly, the NP gene phylogeny showed H6 aquatic isolates with a sister or basal relationship to each of the major reassortments leading to the emergence of novel H5N1 and H9N2 genotypes (Fig. (Fig.3a).3a). The M gene showed a strong host restriction phylogenetic signal among all established lineages, as the current H6 lineage from domestic duck and H9N2 and H5N1 lineages as well as the gene pool lineage were easily distinguishable in the tree (Fig. (Fig.3b).3b). The NS gene phylogeny revealed that multiple genetic lineages from both allele A and allele B cocirculate in the H6 subtype gene pool from aquatic market birds (Fig. (Fig.3c3c).
In summary, phylogenetic analyses of each of the eight gene segments of those H6 viruses revealed that group I viruses with ST339-like gene constellation were persistently prevalent in domestic ducks from 2000 to 2004 (Table (Table1).1). Even though a few reassortant group I viruses with internal gene segments from the gene pool were observed, these novel viruses were sporadically encountered and not established in domestic ducks. The group II viruses were reassortant group I-like viruses with new surface genes introduced as early as 2002. From 2002 to 2005, group I and group II viruses cocirculated in the field and group II viruses became established gradually in the domestic duck hosts from 2005 onward (see Table S1 in the supplemental material). Further, in 2005 a large number of group II viruses incorporated a new NP gene from the gene pool. The group III viruses analyzed were representative of the influenza virus gene pool in this region which clustered with other subtypes of influenza viruses in all internal gene trees (Table (Table1).1). Among all tested viruses, only two isolates belonged to the H6N1/N2 virus lineage of established terrestrial H6 viruses. The majorities of reassortants, or novel genome constellations, were transient and resulted from gene exchange between group I and group II or the gene pool (Table S1). These findings showed that gene interaction between the influenza virus gene pool and those established virus lineages in domestic ducks occurred throughout the surveillance period.
To understand the antigenic properties of group I, group II, and group III H6 viruses, representatives from each of those three H6 lineages were tested by HI using a panel of reference antisera (Table (Table22 ). None of the identified groups reacted well with anti-Teal/HK/W312/97 or anti-Qa/HK/1564/00, representative viruses of H6 viruses established in terrestrial poultry, whereas all groups (group I, group II, and group III) reacted well with anti-Dk/ST/5540/01, representative of group I viruses. Group II viruses were moderately reactive with anti-Dk/ST/14966/01, representative of group II viruses, whereas group I and group III viruses showed low to moderate reactivity compared to the homologous titer. Group III viruses reacted well with anti-Dk/JX/227/03 viruses, but group I and group II were only moderately reactive. This suggests that even though these lineages are phylogenetically distinct, they are similar in their antigenic properties.
It has generally been accepted that ducks and shorebirds are the natural reservoirs of influenza A viruses (1, 6, 19). However, the role of domestic ducks (compared to migratory ducks) in influenza virus ecology, particularly the interaction and prevalence of different subtypes of influenza viruses between migratory and domestic ducks, has not been fully defined. In this study, phylogenetic analysis of viruses isolated in southern China revealed that multiple lineages of H6 subtype viruses cocirculated in the domestic duck population. Our findings demonstrated that the majority of H6 duck isolates belonged to a single H6N2 virus lineage with a single gene constellation (group I), distinct from the natural gene pool, that had become endemic. Our results also revealed the emergence of a novel H6 lineage (group II) as well as the circulation of viruses from the Eurasian natural gene pool (group III) in this host in southern China.
Influenza A viruses belonging to the natural gene pool, represented by viruses from diverse geographical origins, have been frequently detected from domestic ducks in inland provinces. These viruses cannot be distinguished from those viruses detected directly from migratory ducks or sentinel ducks in our surveillance program. This suggests that domestic ducks in this region form a significant part of the natural influenza virus reservoir.
As one of the major poultry types, domestic ducks in China have the largest population size in the world (7). H6N2 virus was one of the most frequently detected subtypes in our surveillance program. Phylogenetic analyses revealed that most H6N2 viruses isolated in Guangdong and Fujian had the same gene constellation, especially their internal gene complex. The long-term presence of this stable genotype indicates that virus behavior in domestic ducks is similar to that seen in terrestrial poultry species in the region (3). Domestic ducks therefore appear to act as part of the natural reservoir along with wild aquatic birds but also as a major poultry type that harbors endemic virus lineages. Even though H6 viruses were also detected from mallards in North America over many years, those viruses were characterized by frequent reassortment and a stable virus genotype of this kind was not recognized (1, 6).
Reassortment or gene exchange between the established H6 virus lineage and the natural gene pool occurred sporadically throughout the surveillance period (see Table S1 in the supplemental material), but few reassortants were persistent or prevalent. It is likely that those transient reassortant viruses have reduced viral fitness or that there may be a fitness advantage for those established internal gene complexes. This is very different from highly pathogenic H5N1 variants that cocirculate and also show genotypic replacement over relatively short periods of time, e.g., 3 years (5, 8, 17, 20).
The lack of reassortment in the aquatic bird H6 lineage is interesting given that H5N1, H9N2, and terrestrial bird H6N1/N2 subtype viruses, which are known to readily reassort with each other, have cocirculated in the poultry populations under surveillance (8, 22, 23). Interestingly, two H6 duck isolates from the gene pool (group III) appear to have donated genes for the generation of the dominant H5N1 variants (genotypes Z and V). So, it is not the established H6N2 virus lineage but the viruses from the influenza virus gene pool in domestic ducks that contributed to the genetic diversity of H5N1 and H9N2 viruses (5, 22, 23).
Although the W312-like H6N1/N2 viruses have been endemic in minor terrestrial poultry in southern China for over 10 years, their transmission to domestic ducks is very limited. Interspecies transmission of H6 virus from domestic ducks to terrestrial poultry was also not recognized. These results highlight the host restriction that is prevalent between domestic ducks and other terrestrial poultry. However, our previous studies showed that domestic duck was the major host in which H5N1 influenza viruses were endemic and that reassortment occurred frequently in this host with subsequent transmission to terrestrial poultry (15). The mechanism responsible for the heterogeneity in virus behavior in precisely the same ecosystem remains unclear.
In this study our data show that virus or virus gene introduction from migratory duck to domestic duck occurred throughout the surveillance period. Even though interspecies transmissions of H6 subtype virus from domestic ducks to terrestrial poultry were not common, virus or gene introduction from domestic ducks to other types of poultry of other virus subtypes occurred frequently, especially in the case of highly pathogenic H5N1 virus. Thus, domestic ducks could be treated as intermediate hosts between the “real gene pool” from migratory ducks and those terrestrial poultry in the whole influenza virus ecosystem. Further investigation and surveillance are required to understand the role of the domestic duck population in facilitating virus interaction and the generation of genetic diversity.
This study was supported by the Li Ka Shing Foundation, the National Institutes of Health (NIAID contract HHSN266200700005C), and the Area of Excellence Scheme of the University Grants Committee (grant AoE/M-12/06) of the Hong Kong SAR Government. G.J.D.S. is supported by a career development award under NIAID contract HHSN266200700005C.
Published ahead of print on 12 May 2010.
†Supplemental material for this article may be found at http://jvi.asm.org/.