Previous Article | Next Article ![]()
Journal of Virology, July 2007, p. 7529-7539, Vol. 81, No. 14
0022-538X/07/$08.00+0 doi:10.1128/JVI.00327-07
Copyright © 2007, American Society for Microbiology. All Rights Reserved.

State Key Laboratory of Emerging Infectious Diseases, Department of Microbiology, The University of Hong Kong, Faculty of Medicine Building, 21 Sassoon Road, Pokfulam, Hong Kong, SAR, China,1 International Institute of Infection and Immunity, Shantou University, Shantou, Guangdong 515031, China,2 Department of Infectious, Parasitic and Immune-Mediated Diseases, Istituto Superiore di Sanità, Viale Regina Elena 299, Rome, Italy,3 Department of Public Health and Animal Pathology, Faculty of Veterinary Medicine, University of Bologna, Ozzano Emilia (BO), Italy,4 Istituto Zooprofilattico Sperimentale of Lombardia and Emilia, Parma, Italy,5 Virology Division, Department of Infectious Diseases, St. Jude Children's Research Hospital, Memphis, Tennessee 380156
Received 14 February 2007/ Accepted 3 May 2007
|
|
|---|
|
|
|---|
The H5N1 "bird flu" incident in Hong Kong, China, in 1997 demonstrated that a "pure" avian influenza virus could be directly introduced into humans and cause infection (9, 10, 41), suggesting a possible pathway for the generation of a new pandemic influenza (31). Genetic analyses suggested that the H5N1 virus causing human infection in Hong Kong, represented by A/Hong Kong/156/97 (H5N1/97), might have been generated by multiple reassortment between A/goose/Guangdong/1/96 (Gs/GD)-like and H9N2 or H6N1 viruses found in terrestrial poultry (19, 22, 50). Following the 1997 outbreak, highly pathogenic avian influenza (HPAI) H5N1 variants with Gs/GD-like hemagglutinin (HA) and neuraminidase (NA) genes became endemic in domestic poultry in southern China (6, 17, 18, 31, 42, 46). Those variants arose through a series of reassortment events to incorporate different gene segments with different origins (17, 18, 31). The endemicity of HPAI H5N1 viruses in poultry in China and its subsequent spread throughout Eurasia and to Africa continue to result in outbreaks in poultry and occasional human infections that has escalated into a new pandemic threat (5, 31, 42). However, the fundamental question of how the Gs/GD-like virus and its variants emerged in southern China has not been completely determined.
Previous studies have demonstrated that, in the case of H5 and H7 subtypes, low-pathogenic avian influenza (LPAI) virus precursors may evolve into HPAI viruses once transmitted to domestic poultry (2, 23). Since 1997, several LPAI H5 and H7 subtype outbreaks have been observed in Europe (2, 4, 13, 26, 35). Influenza surveillance in migratory birds and domestic poultry in Italy revealed that the precursor virus responsible for LPAI outbreaks in domestic poultry originated from migratory birds (4, 11-13). Moreover, a study in northern Europe also suggested that LPAI subtypes of H5 and H7 viruses resident in migratory duck were introduced to poultry and caused HPAI H5N2 and H7N7 outbreaks (35). Since 2003, outbreaks of LPAI H5N2 virus have also occurred in chickens in Japan and Taiwan. However, phylogenetic analyses revealed that the virus from Taiwan was most closely related to an LPAI H5N2 virus that caused multiple outbreaks in Mexico in the mid-1990s and not to Eurasian lineage viruses (8). This virus does not match the natural evolutionary pathway of influenza A virus, thereby implying another mechanism of virus emergence.
Systematic influenza surveillance of domestic poultry in Hong Kong and southern China was conducted from 1976 to 1980 and 2000 onwards, while surveillance of migratory bird populations in southern China has been conducted since 2002 (6, 15, 31, 39-42). These surveillance studies accumulated comprehensive data on the influenza virus gene pool in this region and provide a unique opportunity to explore the possible evolutionary pathways of HPAI Gs/GD-like virus and its variants, i.e., different genotypes. In the present study, LPAI H5 subtype viruses isolated from surveillance of both domestic poultry and migratory birds in southern China, together with some H5 subtype viruses from Italy, were genetically and antigenically characterized. Our findings revealed the dynamic nature of the influenza virus gene pool in Eurasia with repeated transmissions between the eastern and western extremities of the continent. Furthermore, our findings suggest that current H5N2 and H5N3 influenza viruses in domestic birds of southern China were directly derived from those viruses resident in migratory birds, particularly from migratory ducks. Epidemiological and genetic findings also suggest that those LPAI H5 subtype viruses detected in our surveillance have not become established in poultry in southern China. Accumulated evidence suggests that Gs/GD-like virus might be derived from viruses resident in migratory birds along the western Pacific flyway. Some Gs/GD-like variants from poultry appeared to have resulted from further reassortment from viruses in migratory birds. However, there are some H5N1 variants that did not match natural evolutionary pathways, suggesting other origins for those viruses. The current study highlights the necessity of continued systematic influenza surveillance in both domestic and migratory birds in Eurasia to track the emergence and reemergence of novel influenza viruses with pandemic potential.
|
|
|---|
Three different representative groups of LPAI H5 subtype viruses were characterized and analyzed (Tables 1 and 2). These included 12 viruses isolated from domestic poultry in Hong Kong between 1975 and 1980, 19 viruses isolated in southern China since 2000, and eight viruses isolated from poultry and mallards in Italy. To provide further information on the influenza virus gene pool in both migratory birds and domestic ducks of this region, 11 viruses from seven different subtypes isolated from our surveillance were also characterized and analyzed (Table 2).
|
View this table: [in a new window] |
TABLE 1. H5N3/H5N2 viruses sampled in southern China from 1976 to 1980 and 2000 to 2005
|
|
View this table: [in a new window] |
TABLE 2. Additional viruses characterized in this study
|
To visualize similarity between the antigenic reaction patterns of different viruses, numerical analysis of HI titers was conducted using PRIMER version 5.2.9 (PRIMER-E, Plymouth, United Kingdom). Titers of <100 and >12,800 in the HI assay were converted to 0 and 25,600 in this analysis, respectively. The data were square root transformed, and the Bray-Curtis coefficient (3) was used to construct a similarity matrix. Hierarchical agglomerative clustering with group-average linking (43) was conducted, and a dendrogram was produced. Nonmetric multidimensional scaling (28) was also used to produce two- and three-dimensional ordinations over 100 iterations. The two-dimensional configuration with lowest overall stress was presented.
Phylogenetic and molecular analyses. To better understand the origins of the Gs/GD lineage of H5 HPAI, 18 LPAI H5N2/N3/N8 viruses obtained from 2002 to 2005 and 12 LPAI H5N2/N3 viruses isolated during 1976 to 1980 from southern China, together with eight H5N2/N3 viruses from Italian poultry or mallards, were sequenced for each of the eight gene segments (Tables 1 and 2). To provide the background genetic information from a further 10 different subtypes, viruses isolated from migratory ducks in the last several years were also sequenced and phylogenetically analyzed with all available sequence data in public databases (Tables 1 and 2).
RNA extraction, cDNA synthesis and PCR were carried out as previously described (6). Sequence assembly, editing, alignment, and residue analysis were performed as previously described (6, 20). Phylogenetic analysis using MRMODELTEST 2.2 (36), PAUP* 4.0 (44), and MRBAYES 3.1 (24) also was carried out as previously described (6).
Nucleotide sequence accession numbers. The nucleotide sequences obtained in this study are available from GenBank under accession numbers EF597247 to EF597498.
|
|
|---|
Antigenic analysis. Antigenic analysis with MAbs against representative H5 influenza viruses demonstrated a diversity of reaction patterns (Table 3). LPAI H5 viruses isolated from both domestic and wild waterfowl in southern China since 2003 onwards were antigenically conserved and did not differ significantly (greater than fourfold) from each other in the HI assay. Those viruses obtained from both poultry and mallards in Italy also had very similar reactivity patterns. Only Ck/Italy/9097/97 had a significantly lower titer to MAb 15A3 against VNM/1203/04. It is interesting to note that most LPAI H5 viruses isolated from Italy and southern China from the 1970s to the mid-1990s also showed similar reaction patterns (Table 3). In contrast, HPAI H5N1 representative strains showed high antigenic diversity with different reaction patterns, as has been previously observed (6, 42). It was noted that the reaction pattern of Ck/PA/1370/83 was different from those of all Eurasian viruses tested (Table 3).
|
View this table: [in a new window] |
TABLE 3. Antigenic analysis of H5 subtype influenza viruses by hemagglutinin inhibition test
|
![]() View larger version (17K): [in a new window] |
FIG. 1. Numerical analysis of HI titers (see Table 2) by nonmetric multidimensional ordination in two dimensions (A) and using hierarchical agglomerative clustering (B). Colors indicate viruses isolated from Italy (blue), southern China (green), and A/goose/Guangdong/1/96-like (red). Abbreviations: IDN, Indonesia; Mall, mallard; MDk, migratory duck; BHG, bar-hooded goose; QH, Qinghai.
|
![]() View larger version (45K): [in a new window] |
FIG. 2. Phylogenetic relationships of the HA (A) and M (B) genes of representative influenza A viruses isolated in Asia, America, and Europe. Numbers above and below the branch nodes indicate neighbor-joining bootstrap values of 50% and Bayesian posterior probabilities of >95, respectively. Analyses were based on nucleotides 22 to 1032 of the HA gene and 26 to 947 of the M gene. The HA and M gene trees were rooted to A/Japan/305/57 and A/equine/Prague/1/56, respectively. Colors indicate viruses isolated from Italy (blue), southern China (green), and A/goose/Guangdong/1/96 (red). Red branches indicate A/goose/Guangdong/1/96-like highly pathogenic influenza virus and its variants. Bar, 0.1 nucleotide substitution per site. Abbreviations: Am, America; Mall, mallard; MDk, migratory duck; Gf, guinea fowl; SCk, silky chicken; BH goose, bar-hooded goose; QH, Qinghai; Qa, quail; WDk, wild duck.
|
Group 2 consists exclusively of viruses from Western Europe, with the exception of the virus Dk/Chany Lake/9/03 that was isolated in Novosibirsk, Russia in Central Asia (Fig. 2A). In contrast, group 3 contained viruses from the western Pacific flyway isolated from domestic ducks and swans, along with HPAI Gs/GD-like H5N1 viruses that have subsequently spread throughout Eurasia (6, 31, 42). This group therefore seems to match one of the major migratory flyways in this region (47).
In the American lineage, viruses from two major outbreaks and those from migratory birds clustered together (Fig. 2A). It has been noted that a virus from Japan, Dk/Hokkaido/84/02 also joins this lineage, providing an example of gene exchange between the American and Eurasian influenza virus gene pools. Interestingly, a virus that recently caused outbreaks in poultry in Japan in 2005 (Ck/Ibaraki/1/2005) also fell in the American lineage and was most closely related to Ck/Hidalgo/28159-232/94 and Ck/Taiwan/1209/03 that caused outbreaks in Mexico and Taiwan, respectively (Fig. 2A) (8).
Phylogenetic analysis showed that three N2 genes from recent southern China H5N2 isolates (Gs/GY/3799/05, Dk/JX/1286/05, and Dk/JX/3345/05) were closely related to the N2 genes of viruses obtained from mallards in Italy and ducks in Hokkaido, Japan (data not shown). Analysis also showed that the N3 genes from the recent H5N3 viruses isolated in southern China are closely related to H7N3 subtype viruses from domestic and migratory birds in Italy from 2001 to 2003 (data not shown). Therefore, both the HA and NA genes of those H5N2 and H5N3 viruses recently isolated from domestic ducks or migratory birds in southern China were from the Eurasian lineage and appear to be derived from viruses resident in migratory birds. Phylogenetic analysis of N1 NA also revealed that the NA gene of Gs/GD-like virus was most closely related to Dk/Hokkaido/55/95 (H1N1) (data not shown). Another H5N1 variant (Ck/Hebei/718/01) had an N1 that was almost identical to African starling/England-Q/983/79 (H7N1) (unpublished data), which was investigated as a vaccine candidate in China (7, 33, 37).
Phylogenetic analysis of the internal genes. In general, all internal gene trees were divided into the American and Eurasian lineages, and the "early" and "contemporary" sublineages were observed within the Eurasian lineage (Fig. 2B and 3 and data not shown). The internal gene phylogenies further highlight the dynamic nature of the influenza virus gene pool in Eurasia with viruses from the east and west frequently clustering together. For the LPAI H5N2 virus Ck/Ibaraki/1/05, all available gene segments clustered with the Mexico-like HPAI from 1994. It is noteworthy that several Gs/GD-like variants isolated from northern China from 1997 to 2004 consistently clustered within the "early" Eurasian sublineage (Fig. 2B and 3 and data not shown). Most of the LPAI H5 viruses from southern China clustered or grouped together with some viruses from European viruses, including those Italian HPAI and LPAI viruses (Fig. 2B and 3 and data not shown).
![]() View larger version (56K): [in a new window] |
FIG. 3. Phylogenetic relationships of the NP (A), PA (B), and PB2 (C) genes of representative influenza A viruses isolated in Asia, America (Am), and Europe. Numbers above and below the branch nodes indicate neighbor-joining bootstrap values of 50% and Bayesian posterior probabilities of >95, respectively. Analyses of NP, PA, and PB2 genes were based on nucleotides 41 to 984, 1425 to 2135, and 1065 to 2085, respectively. The gene trees were rooted to A/equine/Prague/1/56 (NP, PB2) and B/Lee/40 (PA), respectively. Colors indicate viruses isolated from Italy (blue), southern China (green), and A/goose/Guangdong/1/96 (red). Red branches indicate A/goose/Guangdong/1/96-like highly pathogenic influenza virus and its variants. Bar, 0.01 nucleotide substitution per site. Abbreviations: Am, America; Qa, quail; Mall, mallard; Mdk, migratory duck; Gf, guinea fowl; Sck, silky chicken; BH goose, bar-hooded goose; QH, Qinghai; Ph, pheasant.
|
In the NP gene tree, all recent LPAI H5 viruses from sentry ducks and retail markets in southern China clustered together with viruses isolated directly from migratory ducks at Mai Po Marshes in Hong Kong or Poyang Lake in China (Fig. 3B). All 1970s H5 viruses joined the "early" sublineage, while the Italian viruses were most closely related to other viruses from Europe. Interestingly, Dk/Hokkaido/120/01 solely clustered with three viruses all isolated from migratory or sentry ducks at Poyang Lake (Fig. 3B). These relationships were also observed in the PB2 and PA trees, suggesting that this virus might also have originated in migratory ducks (Fig. 3). The PA gene tree had a phylogenetic topology similar to that of the NP gene tree (Fig. 3B). It is noteworthy that those Gs/GD-like variants clustered together with other viruses from migratory birds and formed different clades within the contemporary Eurasian sublineage.
In the PB2 phylogeny, the Eurasian lineage formed three groups. The first group consisted of some early LPAI H5 viruses from domestic ducks in southern China and Japan, along with recent H5N1 viruses from northern China (Fig. 3C). The other two groups consist of both early lineage viruses and recent isolates, in contrast to the other gene trees where the early viruses almost exclusively formed a single group. Interestingly, the PB2 gene of Gs/GD/1/96 is most closely related to that of an H11N3 virus (WDk/ST/1411/00) that was isolated from mallard at a market in Shantou, China. Whether this bird was farmed or truly migratory is not known. Also these two viruses were then most closely related to Ck/Italy/9097/97, a virus that appeared to have been recently introduced to poultry from migratory birds. Moreover, many Gs/GD-like variants, including Qinghai-like virus, clustered with an LPAI H5 virus (Dk/YN/435/02) that incorporates several gene segments of migratory bird origin. These findings provide further evidence that Gs/GD-like virus and some of its variants may have been introduced to poultry from migratory birds.
To identify the possible source of Gs/GD-like HPAI H5N1 viruses and its variants, we summarized their most phylogenetically closely related viruses for each of the gene segments (Table 4). For Gs/GD-like viruses, three of the gene segments were likely derived from migratory birds, and four of the genes are most closely related to those viruses isolated from ducks in Hokkaido, Japan. Furthermore, some of the variants isolated from our surveillance during 2001 to 2005 in southern China contain gene segments that are most closely related to gene segments from viruses in the migratory bird gene pool (Table 4). It is noteworthy that these analyses confirm that several H5N1 variants isolated from northern China from 1997 to 2004 contained many gene segments belonging to the early Eurasian gene pool identified from viruses isolated in the 1970s and 1980s (27). Moreover, it is surprising that one H5N1 virus (Ck/Hebei/718/01) contained an NA N1 gene with very high homology (98.3%) to that of African starling/England-Q/983/79 (H7N1).
|
View this table: [in a new window] |
TABLE 4. Possible sources of gene segments of highly pathogenic H5N1 influenza viruses isolated in China, 1996 to 2005
|
|
View this table: [in a new window] |
TABLE 5. Connecting peptide of H5 subtype viruses characterized in this study
|
|
|
|---|
Phylogenetic and antigenic analyses demonstrated that LPAI H5 viruses obtained from southern China in recent years were closely related to those H5 viruses isolated from Italy in the mid-1990s (2, 4, 11-14), while different subtypes of influenza viruses isolated from migratory birds in southern China clustered together with either viruses from Hokkaido (Japan), Mongolia, and Siberia, or with viruses from Europe. These findings provide further evidence of the dynamic influenza virus gene pool in this region. Along the western Pacific migratory flyway, the influenza virus gene pool in the domestic waterfowl of southern China has "mixed" longitudinally with viruses isolated from Japan, Mongolia, and Siberia. However, it appears that there has also been "mixing" latitudinally through overlapping migratory flyways, thereby facilitating interaction between the influenza virus gene pool in domestic waterfowl in the eastern and western extremities of the Eurasian continent. This helps to explain the latitudinal spread of the Qinghai-like (clade 2.2) H5N1 virus in the last 2 years, while H5N1 outbreaks in Korea and Japan may represent the longitudinally transmitting pathway (5).
In the present study, the phylogenies of Gs/GD-like H5N1 virus genes either clustered with viruses from migratory birds or with viruses isolated from ducks in Hokkaido, Japan (Table 4). As no systematic surveillance was conducted from 1980 to 1997 in southern China, we failed to identify the direct precursor of the Gs/GD-like virus which emerged in 1996 (50). Even though it is not clear whether those Hokkaido viruses were from domestic or migratory ducks, many of them clustered directly with the viruses isolated from migratory ducks at Poyang Lake, Jiangxi Province, China, and Mai Po Marshes, Hong Kong, since 2002 (Fig. 2 and 3 and data not shown). We therefore speculate that the duck viruses from Hokkaido may also be of migratory bird derivation. Thus, the available findings suggest that Gs/GD-like H5N1 virus may be derived from an LPAI H5N1 virus in migratory waterfowl along the western Pacific migratory flyway. A precedent for such an introduction and subsequent change from LPAI to HPAI has also been observed in Europe (2, 4, 35).
Phylogenetic analysis also showed that those H5N1 variants isolated in southern China from 2001 to 2005 were mainly reassortants between Gs/GD-like virus and virus from migratory birds. Even though there are still some gene segments for which we could not identify the source, current results suggest that these reassortment events very likely occurred in the field. However, those H5N1 variants from northern China in 1997 and afterwards contained almost all internal genes that were obviously from early gene pool viruses (Table 5). The remaining question is to how those old gene segments had been incorporated into Gs/GD-like virus to generate those variants.
Two possibilities could explain this situation, either these data were a result of laboratory contamination or the viruses were isolated in the field through the use of incompletely inactivated poultry vaccines that utilized older strains of LPAI H5 virus. For instance, one virus (Ck/Hebei/718/01) contained an N1 gene that was almost identical to that from African starling/England-Q/983/79 (H7N1). This H7N1 virus, along with Ty/Wisconsin/1/66 (H9N2) and Ty/England/N28/73 (H5N2), were investigated as vaccine candidates in China (33, 37). An H9N2 virus (Ck/Heilongjiang/35/00), isolated in a field in northern China, had almost identical HA and NP genes with Ty/Wisconsin/1/66 (30, 34). Furthermore, genetic analyses of viruses recently isolated from LPAI H5N2 outbreaks in Taiwan (8) and Japan confirmed that those H5N2 viruses were derived from a Mexico-like LPAI H5N2 virus, suggesting that those viruses may have been directly derived from incompletely inactivated vaccine (29). Therefore, it seems likely that those H5N1 variants containing early gene segments were not generated under natural conditions.
Currently, HPAI H5N1 viruses are endemic in poultry in many Eurasian countries (6, 31, 42), and as this article was being prepared, a new outbreak wave was occurring across eastern and southeastern Asia. Genetic analysis of the recent outbreak suggested that the outbreaks in Korea and Japan were caused by BH goose/QH/65/05-like virus. If this is the case, it is not impossible that those Gs/GD-like variants may have successfully become established in the natural reservoirs of influenza viruses, i.e., migratory waterfowls. If so, control of HPAI H5N1 in the short term has become an impossible mission.
This study demonstrates the dynamic nature of the influenza virus gene pool in Eurasia with continuing gene exchanges between the eastern and western parts of the continent. Our findings suggest that the HPAI Gs/GD-like virus were likely derived directly from viruses resident in migratory birds. The failure to identify the source for all gene segments of Gs/GD-like viruses and its variant highlights the need for continued and extensive surveillance in both migratory and domestic populations in larger regions. Such surveillance is crucial for effective pandemic influenza preparedness.
We gratefully acknowledge the excellent technical assistance of Li-Juan Zhang and Justin Bahl and the continual and invaluable collaboration of Marco Carsughi, Antonio Canu, and Luigi Calchetti (WWF Italian Delegation).
Published ahead of print on 16 May 2007. ![]()
|
|
|---|
This article has been cited by other articles:
| |||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
Copyright © 2009 by the American Society for Microbiology. For an alternate route to Journals.ASM.org, visit: http://intl-journals.asm.org | More Info»